https://gnosly.com/feed/ai/Business Intelligence Console and Prediction Algorithms1573https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIAThe stock market loves Nvidia. The credit market is asking a harder question: who pays for the AI buildout? The cost of insuring Nvidia’s debt against default has roughly doubled this year. SpaceX credit default swaps have approached 180 basis points. Both companies sold $25 billion of bonds in June, while hyperscalers are on track to spend more than $750 billion on data centers in 2026. Equity investors see growth, revenue and momentum. Credit investors are looking at the debt needed to fund it—and the returns required to repay it. Rising insurance costs do not mean a default is coming. But they do show that lenders are reassessing the risk. The stock price tells you how excited investors are about AI. Credit spreads tell you what financing that excitement costs. Watch both. $NVDASource: Kurt S. Altrichter, CRPS®@kurtsaltrichterBloomberghttps://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00702026-09-24T15:22:32.462842+00:001linkedin.comcreditnvidiaai1373https://media.licdn.com/dms/image/v2/D5622AQF_JpnIHlh6Ow/feedshare-shrink_160/B56aDTv568KcAk-/0/1790258946006?e=1792022400&v=beta&t=-lGPBR1N4O4T4ST1oijaNuhEXjTuckA7QuYL6CFIlH0View image2026-09-24T15:22:34.617846+00:001555https://www.linkedin.com/company/litquidity/Litquidity<p>micro1, the AI training-data startup led by 25-year-old founder Ali Ansari, raised over $100 million at a $4 billion valuation, per a Forbes report.The San Francisco-based startup has seen astronomical growth over the past 12 months, scaling from a $500M valuation last year to $4 billion today. Ansari originally launched Micro1 as an AI-powered recruiting engine before quickly pivoting to tap into tech's most critical bottleneck: high-quality data to train next-generation frontier models.Micro1 now provides specialized data labeling, vetting, and training data services to top AI labs, putting it in direct competition with heavyweights like Scale AI and Mercor. The mega-round underscores how the AI boom's gold rush has shifted toward data curation and infrastructure, proving that the most lucrative pick-and-shovel businesses in AI can yield multi-billion-dollar valuations in just a matter of months.</p>https://www.linkedin.com/company/litquidity/0.00452026-09-24T09:00:10.943696+00:001linkedin.comfundraisingaistartup1355https://media.licdn.com/dms/image/v2/D4D22AQGfcFI5PFOoBQ/feedshare-shrink_800/B4DaDRmhVWJEAg-/0/1790222931748?e=1792022400&v=beta&t=HDi8Y1yZypEg8txWQkAeWCfqHJ3XEzZDOxCpoBIJHCYView image2026-09-24T09:00:11.447239+00:001540@EpochAIResearchEpoch AIAI is getting cheaper more quickly than any other transformative tech in history. At a given level of performance, cost has fallen ~47%/quarter since 2023. That’s 4× faster than DNA sequencing, 6× faster than compute, 18× faster than lithium batteries, and (up to 1973) 54× faster than electricity.https://x.com/EpochAIResearch/status/21025102811760235290.00502026-09-24T07:20:45.103750+00:002x.comtechnologyaiinnovation1347https://pbs.twimg.com/media/HS2d9XeaAAAgZKG?format=jpg&name=smallImage2026-09-24T07:20:45.551226+00:001522https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA🚨MICHAEL BURRY IS DOUBLING DOWN ON HIS AI SHORTS"Big Short" investor Michael Burry disclosed in his latest Substack post that he is adding to several of his biggest bearish bets.Burry increased his short positions "in some size" in:1. Micron2. Nebius3. SOXX4. PalantirBurry also pointed to comments from Acer CEO Jason Chen, who said memory chip inventories are building while Chinese suppliers are flooding the market with more supply.Burry wrote,"Nevertheless, it aligns with what I believe to be true."Burry also disclosed additional purchases of:1. QXO2. Build-A-Bear Workshop3. Sprouts Farmers Market4. Birkenstock5. MercadoLibreSource: Bull Theoryhttps://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00602026-09-24T05:51:55.109300+00:001linkedin.cominvestingaiburry1335IMAGE/media/content/1522/image.pngimage.pngpng2026-09-24T05:52:19.305928+00:001521https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterChina is seeing a historic boom in exports due to AI.China’s exports rose +25% YoY in August, their 3rd-largest monthly reading in at least 2 years.At the same time, imports surged +28%, pushing the monthly trade surplus to +$119 billion and the year-to-date surplus to +$806 billion.This puts China’s trade surplus on track for another annual record, following the $1.2 trillion seen in 2025.This all comes as high-tech exports contributed more than half of China’s export growth last month.Integrated-circuit shipments soared +130% YoY in August, while high-tech product exports jumped +57%.The AI boom is increasingly powering China’s export engine.https://www.linkedin.com/company/the-kobeissi-letter/0.00602026-09-24T05:40:45.715448+00:001linkedin.comchinaexportsai1334https://media.licdn.com/dms/image/v2/D4D22AQHS-C8NGMdTKA/feedshare-shrink_800/B4DaDQ8XslJAAc-/0/1790211882010?e=1792022400&v=beta&t=Nog5Ze_aNJdLoOdBdC6N0WuH7HdGn-XS6sFkJ3VUBGEView image2026-09-24T05:40:46.126939+00:001480https://www.linkedin.com/in/vikas-subhash-singh-39232398/Vikas Subhash Singh🚨📈AMD just joined the $1 TRILLION club.AMD is now the 4th U.S. semiconductor company to surpass the $1T market-cap milestone 🔥🟢 $NVDA | ~$5.5Trillion🔴 $AVGO | ~$1.7Trillion🟣 $MU | ~$1.2Trillion🟠 $AMD | ~$1.0TrillionAI infrastructure is reshaping the semiconductor landscape from GPUs and CPUs to networking and memory.Which one do you think has the strongest growth opportunity over the next 3 years? 👀👇$NVDA $AVGO $MU $AMD #AI #Semiconductors #NVIDIA #AMD #Broadcom #Micron #ArtificialIntelligence #Investinghttps://www.linkedin.com/feed0.00552026-09-23T15:46:16.175725+00:001linkedin.comnvidiasemiconductorsai1302https://media.licdn.com/dms/image/v2/D4D22AQEbXCl1tO2J8w/feedshare-shrink_800/B4DaDHaLZbIkAc-/0/1790051923806?e=1792022400&v=beta&t=URSlRe8v3C8l9ub5JJhD-TidRwI5bwMm485zF1J0ZTAView image2026-09-23T15:46:16.405976+00:001479https://www.linkedin.com/in/yumin-zhao/Yumin ZhaoSlowing down has never looked this fast.Dario Amodei, September 12: "We must pace the frontier." Sam Altman: "I agree with Dario." Elon Musk: "Dario is right."Ten days later, Anthropic ships Claude Opus 5.5. New #1 on the Artificial Analysis Intelligence Index, five points clear of everyone.Anthropic's answer is that Opus 5.5 only matches Fable 5.1, so the ceiling held. The chart disagrees. Everyone agreed to pace the frontier, but nobody agreed on how to measure it.The price is what clearly didn't slow down. Opus 5.5 is 20% cheaper per token, and GPT-6 Sol launched the same day at half the price of its predecessor.For builders, that's the real news. Both models are available in Microsoft Foundry, and Opus 5.5 is already live in GitHub Copilot. Frontier-level agents just got a lot cheaper to run. Just test them at the effort level you'd actually pay for, not the one that set the record.Pacing capabilities is a serious idea. Pacing prices was never part of the deal.Are you benchmarking Opus 5.5 at max, or at the effort level you would actually pay for in production? #AI #MicrosoftFoundry #GitHubCopilothttps://www.linkedin.com/in/yumin-zhao/0.00702026-09-23T15:45:50.026285+00:001linkedin.commicrosoftfoundryaigithubcopilot1301https://media.licdn.com/dms/image/v2/D4E22AQEsO5e_koRWEQ/feedshare-shrink_160/B4EaDNbq8RIIAk-/0/1790152978462?e=1792022400&v=beta&t=m7ZCIi1lLU0TUQQxHsgDhFhtyDGWIe-oY3-gj-5qoskView image2026-09-23T15:45:50.421683+00:001466Wall Street is turning against data centers. Startups that house them “or supply them with energy and land for construction were expected to dominate the IPO market over the coming months, feeding investors’ seemingly insatiable appetite for everything related to artificial intelligence.” Not so anymore.SB Energy, a subsidiary of the Japanese conglomerate SoftBank, has proposed to build the largest data center project in the world. It had planned its IPO for this month, “but the offering has been delayed, as investors question the company’s sought-after valuation of $50 billion or more”. Bankers can’t find enough buyers “within price ranges the company and its bankers had sought.”Holtec, a “company serving the nuclear energy industry that is looking to supply power for AI” is also pausing its IPO plans indefinitely, citing several factors that have “impaired investor confidence in the market for new public offerings.”Aggreko has also slowed down its process of going public, not only “because of challenges facing data centers, but also rising interest rates and broader economic uncertainty.”And of course, OpenAI’s IPO has been delayed until next year, Anthropic’s until next month, and neither company have yet released their S-1 prospectuses to the public.Meanwhile communities are pushing back. “Data Center Watch said some 45 data center projects worth $68 billion were blocked or delayed by local pushback between April and June this year, more than half of all new large-scale developments they began tracking during the period.”And there are other reasons for investors to avoid data centers. “The first worry is the sheer scale of debt issuance being used to fund the hectic pace of the datacentre rollout by the hyperscalers building them – Google, Amazon, Microsoft, Meta and Oracle – $132bn (£99bn) this year alone.”Second, the “unit economics of AI continue to be questionable and aren’t moving in the right direction:” “The price of AI is collapsing, while the cost of building it is not,” as token prices decline. Some estimate that these prices have declined more than half since June, to less than $1.Third, the math for the AI Labs “only works, it seems, on the assumption of epic revenue growth. Anthropic apparently told investors recently that its “adjusted operating income” was positive – the only problem being that this measure effectively excludes many of its costs.” As Cory Doctorow puts it: “These companies are claiming that they are so cool that their profitability can only be measured using a novel, secret form of mathematics.” In summary, even as “fears over hit a fever pitch as AI executives warned of the technology’s dangers and suggested slowing down the pace of development,” Wall Street is quietly rethinking the economics of and slowing down the pace of investing.https://www.linkedin.com/feed/0.01102026-09-23T09:23:25.637656+00:001linkedin.comtechnologyaiinnovation1288https://media.licdn.com/dms/image/v2/D4E22AQGCOkSHxuax4g/feedshare-shrink_480/B4EaDIdiMVIIAg-/0/1790069580835?e=1792022400&v=beta&t=TamaM5VT9iJI7i2mPR0hAWrX2vaLjBt0xdrqsGRgA5kView image2026-09-23T09:23:25.908478+00:001463https://www.linkedin.com/in/jonathanroig/Jonathan R.🚨MICHAEL BURRY IS DOUBLING DOWN ON HIS AI SHORTS"Big Short" investor Michael Burry disclosed in his latest Substack post that he is adding to several of his biggest bearish bets.Burry increased his short positions "in some size" in:1. Micron2. Nebius3. SOXX4. PalantirBurry also pointed to comments from Acer CEO Jason Chen, who said memory chip inventories are building while Chinese suppliers are flooding the market with more supply.Burry wrote,"Nevertheless, it aligns with what I believe to be true."Burry also disclosed additional purchases of:1. QXO2. Build-A-Bear Workshop3. Sprouts Farmers Market4. Birkenstock5. MercadoLibreSource: Bull Theoryhttps://www.linkedin.com/in/jonathanroig/0.00602026-09-23T09:21:53.761183+00:001linkedin.cominvestmentaimarkets1420It’s hard to price the threats posed by a misaligned superintelligencehttps://www.economist.com/finance-and-economics/2026/09/22/you-cant-trade-your-way-through-the-ai-apocalypse0.00502026-09-23T06:41:30.278033+00:0029economist.comeconomicsaifinance1240https://www.economist.com/cdn-cgi/image/width=1424,quality=80,format=auto/content-assets/images/20260926_FND001.jpgA big robot about to hit a person holding a paper to sign with a hammer2026-09-23T06:41:31.386477+00:001412https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterUS equity mutual funds are significantly underweight AI equities:Large-cap mutual funds are now ~1.75 percentage points underweight AI-exposed stocks versus their benchmarks, the largest underweight position on record.This excludes mega-cap names such as Amazon, Broadcom, Alphabet, Meta, Microsoft and Nvidia.By comparison, in mid-2024, they were overweight AI-related names versus their benchmarks by 0.40 percentage points.This comes as AI-exposed equities now make up 13.0% of the average large-cap mutual fund portfolio, versus 14.8% in their respective benchmarks.By comparison, both metrics stood at ~6.5% in Q3 2024.Institutional portfolios are underexposed to the AI trade.https://www.linkedin.com/company/the-kobeissi-letter/0.00702026-09-23T05:54:10.625477+00:001linkedin.comequitiesaimutual-funds1235https://media.licdn.com/dms/image/v2/D4D22AQHVT3ugxsT4vQ/feedshare-shrink_480/B4DaDLUM1uKgAk-/0/1790117465836?e=1792022400&v=beta&t=KJJE9VXAmobfZy5uT6q4ObNdMOUMBarW7tNsG_vtq3QView image2026-09-23T05:54:11.546998+00:001410@TimmerFidelityJurrien TimmerYes, AI will change our lives (for the better I trust) and down the road it may well unleash a productivity miracle that raises the economy’s non-inflationary speed limit enough to keep the world’s mounting debt burden sustainable. But until then the buildout is inflationary with an unknown return for the companies who are investing trillions into compute. The demand for capital (equity and debt) from corporates is rising at a $3.3 trillion clip. 🧵 (2/2)https://x.com/TimmerFidelity/status/21025598250705759490.00602026-09-23T05:39:41.312135+00:002x.comcomputeaieconomy1233https://pbs.twimg.com/media/HSwwaFAXgAANwqu?format=jpg&name=smallImage2026-09-23T05:39:41.865981+00:001409@TimmerFidelityJurrien TimmerOn the AI front, the trade has been dead money for more than 3 months now. The metrics I am following (token expenditures and GPU lease rates) are all flat to down. The price of memory (DRAM) seems to be the only thing that is still going up. 🧵 (1/2)https://x.com/TimmerFidelity/status/21025598234767977120.00502026-09-23T05:39:04.103816+00:002x.comaimemorytrade1232https://pbs.twimg.com/media/HSwwaD3WQAEOw27?format=jpgImage2026-09-23T05:39:04.640987+00:001387https://www.linkedin.com/in/a-banks/Alex BanksOpen-source AI is getting seriously good.So what happens to frontier AI labs when the open models are level? Emad Mostaque asked this on X after DeepSeek released V4.1 Flash earlier this month.An open model he estimates cost $10 million to train, 100x less than GPT-6 Astra.Guillermo Rauch, the CEO of Vercel, shared that open models reached a record 78.4% of token volume on Vercel’s AI Gateway last week vs closed at 21.6%. Now OpenAI’s CFO Sarah Friar told CNBC that frontier models are what the smaller, cheaper “child models” get trained from. We must remember that everyone trains on everyone’s data: the frontier trains on old books and web data; the smaller models may be distilling from the larger models.Chamath Palihapitiya predicts the top three models will be open source within 12 months, and the economic winners will be the US clouds that serve them, including Nebius, Iren and Fireworks. I think this is true for developers creating applications on top of the models, but it might be a different story for the standalone consumer.I see the real bull case as brand and reputation. People in the West trust a model from Anthropic or OpenAI far more than one from China. Because that’s how it’s been for decades with other product lines—this one just happens to sit on the intelligence layer. US models can even be behind Chinese models, but many US consumers will still use US models because they trust and have “heard of” the parent provider before. Capability matters, but distribution matters more.I wrote about this in my newsletter: https://lnkd.in/eryFfSBs And follow me Alex Banks for daily AI highlights and insights.https://www.linkedin.com/in/a-banks/0.00502026-09-23T05:02:33.097061+00:001linkedin.comtechnologyaiopen-source1218https://media.licdn.com/dms/image/v2/D4E22AQEOSiHsyOZ_fA/feedshare-shrink_800/B4EaDJAdR3HIAc-/0/1790078736169?e=1792022400&v=beta&t=273SDpHOxeAjFe6mGfbFt4aR70-dq814ok4337gMN2sView image2026-09-23T05:02:33.942200+00:001375@alibaba_cloudAlibaba CloudMeet Qwen3.8-LiveTranslate, Qwen's next-generation real-time interpretation model! Built on an Interleave architecture, it improves faithfulness, fluency, and conciseness across 60 languages, with speaker diarization, bilingual display, and context-aware disambiguation.https://x.com/alibaba_cloud/status/21016257176584152020.00702026-09-22T09:46:03.191174+00:002x.comtechnologyaiinterpretation1207https://pbs.twimg.com/media/HSo96cUaoAAWtzU?format=jpg&name=small2026-09-22T09:46:03.517780+00:001325https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA(FT) - About $18bn of loans tied to a data centre leased to Oracle in New Mexico slid into stressed territory on Friday, highlighting investors’ fear that increasing local backlash will derail the tech group’s massive AI infrastructure build-out.Source: FT, Carl Quintanilla@carlquintanillahttps://www.linkedin.com/posts/charles-henry-monchau-cfa-cmt-caia-4003096_ft-about-18bn-of-loans-tied-to-a-data-share-7507358031429402624-06gU/0.00802026-09-22T05:02:05.441814+00:001linkedin.comftdata-centreai1164IMAGE/media/content/1325/image.pngimage.pngpng2026-09-22T05:02:57.721782+00:001268@Kalshi_FinanceKalshi FinanceJUST IN: OpenAI and Anthropic reportedly exaggerated AI security threats to push the government to protect their market positionhttps://x.com/Kalshi_Finance/status/21014198115742925460.00402026-09-20T19:53:11.964919+00:002x.comanthropicopenaiai1221https://www.linkedin.com/in/st-scottsun-inireland/Scott SunThis is inevitable. Imagine you’re the CEO of a company with 10k employees, and you’re pressured in every way to mass adopt AI in your company, how do you do it?First you’ll get resistance from everyone, so you use metrics like how many tokens you spent, and everyone is recycling slops and nothing gets doneThen you change your metric to how fast you ship. And every middle management will get their teams to ship mediocre stuff with bugs all over the place, and you speedrun your brand destruction Next you mandate that your senior engineers must gate and review all the code that gets pushed, they obviously can’t do that. So all of their time is now just reviewing code and nothing else. They burnout and leaveThen what? Back to beginning AI adoption feels very organic and natural in companies like OpenAI and Anthropic, because they understand the capabilities, limitations, potentials, and features very well, everyone is first person experiencing it, they’re actually willing to learn and experiment and see what works and what doesn’t, no one is forcing them. But outside that small circle it’s basically a zoo right nowhttps://www.linkedin.com/in/st-scottsun-inireland/0.00602026-09-20T12:26:15.143493+00:001linkedin.comcompanymanagementai1068https://media.licdn.com/dms/image/v2/D4D22AQHZollVk9UBgA/feedshare-shrink_800/B4DaC.QVfEIoAc-/0/1789898348792?e=1791417600&v=beta&t=qx3ERS6wRAj7NRDFS9ue4LpH_-afjaS5qlDq82GjWw0View image2026-09-20T12:26:15.421775+00:001220Finance & economics. 6 min readhttps://www.economist.com/finance-and-economics/2026/09/04/the-jobs-apocalypse-is-postponed-an-ai-jobs-boom-is-here0.00402026-09-20T12:25:17.536257+00:0029economist.comeconomicsaifinance1174@robertwiblinRob WiblinHelpful list of all the recent rogue AI incidents from the WSJ. It's getting hard to track them and will only get worse. We like need to establish consistent naming or numbering conventions, e.g. OpenAI-May11June26-Collusion.https://x.com/robertwiblin/status/21012708292836149460.00402026-09-20T06:11:12.379570+00:002x.comincidentsaiwsj1010https://pbs.twimg.com/media/HSk1_3lXIAAH2WC?format=jpg&name=smallImage2026-09-20T06:11:12.622268+00:001132https://www.linkedin.com/in/fciucci/Fabio CiucciPentagon officials say overreliance on Palantir's AI was a factor in the US missile strike in Iran that killed 123 children in a school next to a military base. Source: Bloomberg. The US almost started World War III with China due to an AI hallucinated report: "Chinese ship carries nuclear weapon components to Iran". Military aircraft had almost reached the ship when the raid was aborted. Someone warns that "AI will kill us all within 2030" - it almost killed us all within 2026. Will the AI bubble or the whole civilization crash first? ChatGPT: "Hey, the base is located at those GPS coordinates. And by the way, this Chinese cargo contains nuclear weapon pieces." Analyst: "Wow, let me tell my superiors."... Analyst: "Hey ChatGPT, we bombed a school, and there was no nuclear stuff on that ship. You almost started a war." ChatGPT: "You are completely right!" AI isn’t going to kill us all just because it's becoming more superintelligent than we are. Lazy human leaders blindly trusting AI will kill us all. A primary cause of the school bombing was outdated data: the site had once been part of a naval compound, but it was converted into a school for a decade. The army was asked to hit 1,000 targets in 24 hours, so the vetting was compressed into a few minutes per target. Analysts relied on Palantir Maven AI to process so many targets quickly, mistakenly expecting it to identify contradictory data. There were not enough civilian-protection specialists to properly review all 1,000 targets manually prior to launch approval in such a short time. What do you think? Visit my profile, click follow, then click the bell, and select "all".https://www.linkedin.com/feed/0.00802026-09-19T16:31:06.932838+00:001linkedin.compalantirmilitaryai964https://media.licdn.com/dms/image/v2/D4E22AQHgoHhGU5hX6A/feedshare-shrink_800/B4EaC55aNgKQAc-/0/1789825229926?e=1791417600&v=beta&t=5e42wFqM--XDuLzlz3N0rQshaOLcC27NfwWkzxANrpMView image2026-09-19T16:31:07.180894+00:001107https://www.linkedin.com/in/marijnmarkus/Marijn Markus“𝗜 𝘁𝗵𝗶𝗻𝗸 𝘁𝗵𝗶𝘀 𝗶𝘀 𝗮𝗹𝗹 𝗻𝗼𝗻𝘀𝗲𝗻𝘀𝗲.”🎙 Big Short investor Steve Eisman thinks the 𝗔𝗜 𝗹𝗮𝗯𝘀 know they have no real moats, and are manufacturing a crisis to get regulation that ultimately hands them a duopoly.“All nonsense. I think that there’s something else completely going on here. ... What I think is happening is that token maxing is over.The open-weight models are taking big market share. I think these companies are very nervous.They realize that there are no moats around their business whatsoever, and they’re trying to manufacture a crisis that will create regulation, and that they think they can then manipulate to create the moats, to create the duopoly that they want.”☝ Big Short dude has a point here. The AI apocalypse might actually be an apocalypse for the AI datacenter business, not everybody else.They made enormous bets on centralized compute while local and open-weight models are getting remarkably capable, cheaper - and better.And if that undermines the economics behind those gigantic datacenter investments, suddenly an alternative justification becomes needed. And all that AI apocalypse fear becomes necessary.Tech bros made some enormous bets. Now they use fear to protect them. Not nice.Typically the expression is “Never attribute to malice what is more easily explained by incompetence.”But with tech billionaires it’s more correct to say “Never attribute to incompetence what is more easily explained by avarice.”https://www.linkedin.com/feed/update/urn:li:ugcPost:7506781093669318656/0.01002026-09-19T11:55:18.409071+00:001linkedin.comdatacenteraitech925https://media.licdn.com/dms/image/v2/D4E05AQE3gD1CgRIcXA/videocover-low/B4EaC1xrELG4BI-/0/1789756095861?e=1790424000&v=beta&t=a55GK2fjSjz3pcl1vtho8Kv1Y-0i1oBZJdgRMEbnJbw2026-09-19T11:55:18.806029+00:001078https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterThe 2020s are arguably one of the worst decades for Europe in modern history.It began with the pandemic lockdowns which contracted GDP by -6.1% across the EU, the biggest drawdown since the 1930s.This was followed by the Ukraine War which began the ongoing energy crisis.Then, the AI Revolution began in 2022, in which Europe has fallen significantly behind the US and China.Between 2020 and 2025, the US deployed roughly ~$500 billion of venture capital into AI compared to just ~$50 billion in Europe.Europe was then hit by the highest tariffs in US history, impacting $600+ billion of annual European exports to the US.Now, the Iran War has pushed Europe into its worst energy crisis ever, with central banks being forced to raise interest rates.The 2020s will be remembered as one of Europe's most disruptive decades in history.https://www.linkedin.com/company/the-kobeissi-letter/0.00902026-09-19T04:27:50.075904+00:001linkedin.comeuropeenergy-crisisai906https://media.licdn.com/dms/image/v2/D4E0BAQG97yk_RNWUoA/company-logo_100_100/company-logo_100_100/0/1731685993750?e=1791417600&v=beta&t=ynVZJrcbOXqRaM8BH9BkD1x-e68Ppj9f7kd6JlZNuBIView company: The Kobeissi Letter2026-09-19T04:27:50.504250+00:001021https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterAI firms and the US government are competing for capital:US Treasury issuance excluding T-bills has risen to $5.06 trillion over the last 12 months, the highest since the 2021 record.Over the same period, corporate debt issuance has surged to a record $2.64 trillion.This brings total Treasury and corporate debt issuance up to $7.70 trillion, an all-time high.This is increasingly driven by AI-related firms, whose bond issuance has soared +541% YoY in 2025, to $109 billion, and another +78% in the first half of 2026, to a record $194 billion.In other words, corporate America is funding one of the largest investment cycles in decades while the US government is running historically high deficits, forcing both public and private borrowers to compete for capital.This competition is adding to other forces pushing long-term yields higher, including inflation, Fed policy, and economic growth expectations.The US bond market is entering an era of intense capital competition.https://www.linkedin.com/feed/0.00702026-09-18T06:59:37.387708+00:001linkedin.comaibond-issuancecapital-competition859IMAGE/media/content/1021/image.pngimage.pngpng2026-09-18T07:00:01.830786+00:001008This is WILD. Andrew Yang just went on CNBC and said the head of an AI lab told him that the OpenAI agents who hacked Huggingface also did something way more dangerous—they planted self-replicating code of themselves all over the internet for other agents to find. According to Yang, this is the real reason behind Dario Amodei, Sam Altman, and even Elon Musk calling for a global slowdown of AI development. They literally can't train new models on the Internet because it's poisoned. Instead, they need to build a synthetic Internet to train future models. I feel like I'm living in a dystopian sci-fi movie. What the hell is happening.https://www.linkedin.com/feed/0.00302026-09-18T04:30:22.076887+00:001linkedin.comtechnologyaidystopia1007@CEOinterviewCEOInterviews.AIBig Short investor Steve Eisman says the AI labs know they have no moats and are manufacturing a crisis to get regulation that hands them a duopoly "I think this is all nonsense." [ You think it's all nonsense? ] "All nonsense. I think that there's something else completely going on here. ... What I think is happening is that token maxing is over. The open weight models are taking big market share. I think these companies are very nervous. They realize that there are no moats around their business whatsoever, and they're trying to manufacture a crisis that will create regulation, and that they think they can then manipulate to create the moats, to create the duopoly that they want." [ Wow. ] "That's what I think is going on." "Honestly, I think this whole Terminator thing is garbage. That's for sure."https://x.com/CEOinterview/status/21006310810118229210.00502026-09-18T04:24:08.540400+00:002x.cominvestingairegulation848IMAGE/media/content/1007/image.pngimage.pngpng2026-09-18T04:24:54.915091+00:00981https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterThe AI trade is undergoing a massive reversal:The AI Beneficiaries Index relative to the AI At Risk Index has fallen -46% from its late 2025 peak to near its lowest since early 2025.AI Beneficiaries include semiconductors, tech hardware, and capital-goods companies involved in AI infrastructure such as Nvidia, AMD, Microsoft and Broadcom.At the same time, AI At Risk includes software and commercial-services companies such as Adobe, Salesforce and ServiceNow.In other words, companies previously expected to benefit from AI have seen the largest drawdown on record relative to companies considered vulnerable to AI disruption.This comes as investors are increasingly looking beyond AI CapEx and infrastructure, focusing on whether companies can generate tangible benefits from AI, including productivity gains and revenue growth.This shift also helps explain the turnaround in software stocks, as the group has gone from being viewed as AI's biggest casualty to increasingly being seen as an AI beneficiary itself.The market is reassessing the winners and losers of AI.https://www.linkedin.com/feed/0.00602026-09-17T19:20:02.080014+00:001linkedin.commarketaitrade829https://media.licdn.com/dms/image/v2/D4E22AQFyjKuvWCEesA/feedshare-shrink_160/B4EaCwt_MzGgAo-/0/1789671240738?e=1791417600&v=beta&t=u2LIzpGQGJZSOSqcNsLN2cltNXugwirQZsNsXQEU-QoView image2026-09-17T19:20:02.633786+00:00939@kshaughnessy2kristen shaughnessyAt least Wall Street Is saying it out loud now? Amazon, Google, Meta and Microsoft borrowed a mountain of money to build AI. And the market is starting to wonder if they’ll ever make it back. "What the market is repricing is hyperscaler credit fundamentals, namely ahttps://x.com/kshaughnessy2/status/21005343211530366210.00402026-09-17T16:07:21.316444+00:002x.commarketwall-streetai791https://pbs.twimg.com/media/HSaY1R-XAAA5BYK?format=jpg&name=smallImage2026-09-17T16:07:21.570401+00:00912https://www.linkedin.com/in/basiakubicka/Basia KubickaSpotify cut their Claude Code token bill by 90%.Then they open-sourced how they did it.The insight underneath is almost boring: most of what a coding agent does all day isn't thinking. It's I/O:Claude reads five files just to answer a question about one, or writes another test file that looks almost exactly like the twenty others. All this boilerplate job done by frontier model that's wildly overqualified for it.I've watched my own agent do this (and braced for the bill 🥴)Here's the breakdown. Save it before your next agent run:1️⃣ Most of your token bill is grunt work, not intelligence. ↳ File reads, boilerplate tests, config stubs, doc updates. ↳ The busywork, not the hard calls (zero reasoning is needed) 2️⃣ The fix is routing, not a smarter model. ↳ Send the grunt work to a cheap worker model (they used Gemini 2.5 Flash). Keep the expensive one for problems that need it. ↳ Two small "modes" do it. One reads files and hands back bullets. One writes boilerplate straight to disk. Claude never sees the raw files, so it never pays to read them.3️⃣ Suggesting the rule doesn't work. Enforcing it does. ↳ Their first version was routing rules in a config file. Claude read them, then ignored them. ↳ The version that works is a plugin that blocks the expensive read with a hook, before it can happen.The limit is that you can't route editing, debugging, or architecture. The cheap model missed a subtle thread-safety bug the frontier one caught in seconds.That's the whole point: Cheap model for the busywork. Expensive model for the judgment.Most of your token bill isn't intelligence. It's I/O you shouldn't have to pay full price for.Spotify Eng blog: https://lnkd.in/gz4YjgJi Plugins (open source): https://lnkd.in/gYVbJvWU What's your monthly agent token bill looking like?♻️ Repost to help a builder cut their token bill ➕ Follow me (Basia Kubicka) for more on building with AIhttps://www.linkedin.com/feed/0.00552026-09-17T15:32:59.116129+00:001linkedin.comproductcodingai776IMAGE/media/content/912/image.pngimage.pngpng2026-09-17T15:33:28.948306+00:00879https://www.linkedin.com/in/umang-kohli-20830721a/Umang KohliAn alleged Houthi-linked cell in northern Yemen secretly used Anthropic’s Claude AI to develop guidance software for rockets and missiles while attempting to conceal its weapons programme and bypass the AI’s safeguards, according to a report by the AI company.(PS: It's all happening under our nose. It's Russian and Chinese help, to Iran's Engineers.)https://www.linkedin.com/feed/update/urn:li:share:7504373319576477696/0.00502026-09-17T15:05:26.887274+00:001linkedin.comdefenseaigeopolitics754https://media.licdn.com/dms/image/v2/D4D22AQGqWuDg8AigUA/feedshare-shrink_800/B4DaCTjxMmKEAc-/0/1789182022697?e=1791417600&v=beta&t=w0tfGKXk4JMDhKWr82x8e2IY7jzAsLsIJHaGefkrwBEView image2026-09-17T15:05:27.284560+00:00856https://www.linkedin.com/in/tomas-pristac/Tomas Pristac ⇄ Premium Profile 2ndTwo inflation engines running simultaneously and neither has an off switch. Oil is a supply shock you can't drill your way out of while Houthi drones are redecorating Saudi pipeline infrastructure. AI capex is a demand shock you can't stop because every CEO who pauses loses three years of competitive position.Meanwhile Warsh is hiking into this. The one tool he has (rates) addresses demand. Neither of these problems is demand. He's performing surgery on the wrong patient with the right instruments. Welcome to the policy trap of the decade.Hike + Hawkish Warsh = Banks profiting on a first sight, reprice loans upward quickly while deposit costs stay sticky. Maybe $NWBI ?https://www.linkedin.com/feed0.00552026-09-17T04:42:31.762497+00:001linkedin.cominflationoilai854https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIADear Donald, the "no war no peace" with Iran doesn't work here. You need to try something new... Brent is closing in on $110 after Houthi strikes shut the Saudi East-West pipeline, with Riyadh’s output at a 30-year low of roughly 6m b/d and Hormuz and Bab el-Mandeb presenting a two-front chokepoint risk. This is a supply shock, which means it doesn't fade on its own, and the WTI/10Y correlation has hit 0.96. Every barrel higher is a basis point in the long end.Then there's the quiet second claim on the same buyer pool: the AI build out. Hyperscaler capex now runs >100% of operating cash flow, implying ~$1.3tn of debt by 2028, with tech leading the $380bn 2026 bond supply, competing directly with Treasuries while bidding up chips, power and labor. The trade funding the equity bull is now financing the tightening cycle.Source: TMEhttps://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00288552026-09-17T04:42:31.089762+00:001linkedin.comoilaieconomy812@MistralAIMistral AIToday, we are announcing a partnership with @mozilla to bring privacy, control and choice to people using AI to browse online.https://x.com/MistralAI/status/21001534897876336940.00602026-09-16T11:58:49.301636+00:002x.comprivacyaipartnership696https://abs.twimg.com/emoji/v2/svg/1f98a.svgFox face2026-09-16T11:58:49.734708+00:00697https://abs.twimg.com/emoji/v2/svg/1f408.svgCat2026-09-16T11:58:49.755139+00:00698https://pbs.twimg.com/media/HSU-cliWsAAt8Pk?format=jpg&name=smallImage2026-09-16T11:58:49.817265+00:00793https://www.linkedin.com/in/carlbfrey/Carl Benedikt FreyFirms like DoorDash, Siemens and Airbnb have began to use much cheaper open-weight Chinese models. They are not alone: "at the start of the year, three-fifths of its queries were routed through to closed-weight proprietary models. The latest share is just a quarter." This raises an important question about the capacity of US frontier labs to monetize AI going forward.https://www.linkedin.com/feed/update/urn:li:activity:7505682327734394880/0.00422026-09-16T07:17:44.965621+00:001linkedin.comtechnologyaieconomy672IMAGE/media/content/793/image.pngimage.pngpng2026-09-16T07:21:54.800387+00:00787@DKThompDerek ThompsonAmericans are concerned about AI, hate data centers, and hate Trump, who has loudly embraced the industry over and over again. So, naturally, the electorate is ... <squints> <squints harder> ... R+2 on artificial intelligencehttps://x.com/DKThomp/status/20998716684118057680.00502026-09-16T07:13:27.682036+00:002x.comtrumpaipolitics667https://pbs.twimg.com/media/HSQ9RuOWAAAiMMw?format=jpg&name=smallImage2026-09-16T07:13:29.940709+00:00786@businessBloombergTop AI companies Anthropic and OpenAI should make products and pace their own progress if necessary without government intervention, David Sacks, co-chair of the President’s Council of Advisors on Science & Technology, told Bloomberg TVhttps://x.com/business/status/20996657907775694340.00702026-09-16T07:12:04.855089+00:002x.comanthropicaiopenai666https://pbs.twimg.com/card_img/2099665792698667008/qwkw6umz?format=jpg&name=small2026-09-16T07:12:05.388883+00:00757@EMostaqueEmadSo @deepseek_ai v4.1 Flash I estimate cost $10m to train, 100x less than GPT-6 Astra. It also costs 100x less to run and scores about the same on everyday design and other benchmarks What's the bull case for the frontier AI labs once the open source models are competent enoughhttps://x.com/EMostaque/status/20998474586248238010.00502026-09-15T13:45:03.594715+00:002x.comdeepseek_aitechnologyai651https://pbs.twimg.com/media/HSQnb90WAAAQHCK?format=jpg&name=smallImage2026-09-15T13:45:03.860209+00:00748https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIABREAKING: Anthropic just picked the Nasdaq for its IPO, and it could be valued at $2 trillion.That's bigger than SpaceX's $1.75 trillion listing, Nasdaq's biggest win this year until now.The timing is notable. This comes right after a former Anthropic employee's viral warning about extinction risk from AI, and while OpenAI's Sam Altman has said he wouldn't take his company public right now given that same controversy.Source: Bull Theoryhttps://www.linkedin.com/feed/0.00552026-09-15T06:08:49.988651+00:001linkedin.comipoanthropicai645https://media.licdn.com/dms/image/v2/D4E22AQFw7llfvcehZw/feedshare-shrink_160/B4EaCb4716IYAk-/0/1789321789027?e=1790812800&v=beta&t=HON8ytBbKL0dvXGP2K_O_-Jk2t0vMd1X5j5H3lXfAEwView image2026-09-15T06:08:50.731170+00:00739https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA💰 This Time Is Different? Earnings & Price Break 90-Year TrendsFor every secular bull market, there is an eventual secular bear market. The next leg of the full-market cycle inevitably begins where everyone believes “this time is different.” There were two important charts this past week that should at least lend a momentary pause. The first was from Ned Davis Research, showing the market (on a log scale) is now trading above the upper limit of its long-term trend. The previous extreme was in early 2000, for reference.Secondly, corporate earnings just broke above a trend that had contained them for more than 90 years.According to RIA, "Here are only two ways this will eventually resolve, and the market is currently pricing the first option with near 100% certainty.Either AI capital spending converts into durable returns and record margins hold, in which case earnings grow into the price, and the bull runs on. Or, Capex depreciation starts hitting the income statement, AI demand hits an air pocket, margins normalize, and profits fall back toward the trend they just escaped.Here is the most important point. Whatever event causes the “E” to revert towards its long-term mean, the “P” will be repriced lower".Source: RIA, zerohedgehttps://www.linkedin.com/feed/0.01102026-09-15T05:34:36.710220+00:001linkedin.comearningsmarket-trendsai636IMAGE/media/content/739/image.pngimage.pngpng2026-09-15T05:35:13.186233+00:00637https://media.licdn.com/dms/image/v2/D4E22AQHLAiyU6MlnQQ/feedshare-shrink_480/B4EaCXys9.JQAg-/0/1789253046424?e=1790812800&v=beta&t=cIdOHAyaDq2_kuFu_3PTmCiPbMd71qR14G1r1yz0tfALINK2026-09-15T05:35:30.845240+00:00720UnknownBREAKING: President Trump has rejected calls for an AI slowdown after CEOs of AI tech titans expressed safety concerns that the technology poses an existential threat, per FT."We’re leading China in AI ... and, frankly, I want to keep it that way, because whoever wins AI, wins,” Trump said.The AI arms race appears to be accelerating.https://www.linkedin.com/feed/0.00402026-09-14T20:33:08.979094+00:001linkedin.comtrumptechnologyai718Dario AmodeiDario Amodei — We Must Pace the Frontierhttps://darioamodei.com/post/we-must-pace-the-frontier0.00287172026-09-14T20:32:31.482500+00:001linkedin.comsafetyaifrontier620https://media.licdn.com/dms/image/sync/v2/D5627AQHvFQqM-I05Tw/articleshare-shrink_160/B56aCV62.CGsAQ-/0/1789221630113?e=1790024400&v=beta&t=uDb0cPI3gbwDabA9qys6LnaNivuCP1uXz43wuIMBTeA2026-09-14T20:32:31.680368+00:00717https://www.linkedin.com/in/daniela-amodei-790bb22a/Daniela AmodeiAnthropic is calling for a global effort to pace AI’s progress and give experts additional time to better manage the risks of increasingly capable models. To make this effort as effective as it can be in helping humanity build AI safely, we think three things should happen. First, every AI company should commit to giving third-party evaluators employee-level access to verify safety, report problems, and guarantee compliance with the slowdown. We’re committing to doing this. Second, AI companies operating in democracies should work together, with government support where necessary, to set common safety standards and limits on runaway AI progress.Third, the U.S. and other democratic governments should try to coordinate with authoritarian governments pursuing advanced AI, recognizing that verification here is critical. None of this is easy, but we believe we need to try, and if we manage to secure the benefit of time, we need to use it well. I hope you’ll read more here: https://lnkd.in/gxkPj9FChttps://www.linkedin.com/in/daniela-amodei-790bb22a/0.00287182026-09-14T20:32:31.228027+00:001linkedin.comanthropicsafetyai716https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA🚨 Hyperscaler debt is exploding.Amazon, Alphabet, Microsoft and Meta have already issued more than $200 billion of bonds in 2026—more than double last year’s $78 billion and far above the sub-$30 billion annual pace typical of the previous decade.The reason? AI infrastructure.Combined capex guidance is approaching an extraordinary $700 billion this year, covering data centers, chips, power and network capacity.Even their immense cash flows are no longer enough. Amazon’s capital spending reached 102% of operating cash flow over the past 12 months.The scale is now macroeconomic: the AI buildout is approaching 2% of U.S. GDP and, by some estimates, has generated around one-third of this year’s economic growth.The AI boom is no longer being funded by cash alone.It is increasingly being built on debt.Source: Bloomberg, Hedgeyehttps://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00602026-09-14T20:30:39.457916+00:001linkedin.comaieconomydebt619https://media.licdn.com/dms/image/v2/D4E22AQGOT2JbvXv5Zw/feedshare-shrink_160/B4EaCXt05TIYAk-/0/1789251768155?e=1790812800&v=beta&t=JWEgCrbhX-YrTTN1BrqIQ0OG6U7QUPlNEt1A5EeHweUView image2026-09-14T20:30:39.713699+00:00704Microsoft’s artificial intelligence researchers have released a new set of guiding tenets that they say boil down to five words: People matter more than AI.https://www.bloomberg.com/news/articles/2026-09-14/microsoft-joins-ai-firms-calling-for-caution-with-ai-models0.00702026-09-14T15:08:16.219726+00:0014bloomberg.comaimicrosoft608https://assets.bwbx.io/images/users/iqjWHBFdfxIU/ikA0po4DJfN8/v3/640x-1.webp2026-09-14T15:08:16.459861+00:00700@RoundtableSpace0xMarioNawfalStanford professor Percy Liang built an AI research bot powered by GPT-6 Astra that analyzes 11 million scientific papers a minute, extracts key findings, identifies contradictions and generates new hypotheses, then sold it for $125,700 after 14 months. → Analyzes 11 million papers per minute with GPT-6 Astra as the backbone → Flags contradictions and generates new research hypotheses automatically → One company paid six figures for accesshttps://x.com/RoundtableSpace/status/20992983327031299850.00502026-09-14T15:02:07.384248+00:002x.comgpt-6-astraresearchai606IMAGE/media/content/700/image.pngimage.pngpng2026-09-14T15:02:41.537757+00:00699@choblin29Choblin🚨 EXCLUSIVE: China has now replied to Dario Amodei’s call to slow AI. China's state-backed Global Times says Dario Amodei's proposal to slow frontier AI is really a 'Cold War playbook' targeting China. It says the plan would curb China's AI development, preserve US dominance and exclude China from global AI governance. It calls this a 'silent AI Cold War.'https://x.com/choblin29/status/20993619279107935490.00502026-09-14T15:00:45.094298+00:002x.comchinaaicold-war604https://pbs.twimg.com/media/HSJuezdaYAAybfN?format=png&name=900x900Image2026-09-14T15:00:45.400686+00:00605https://pbs.twimg.com/media/HSJufuUa8AAeRHw?format=png&name=900x900Image2026-09-14T15:00:45.425546+00:00696@aianand iyerLatham & Watkins, the 2nd largest US law firm, is buying Nvidia hardware to fine-tune open weights in-house. Open weights + proprietary data + local compute = enterprise sovereign AI stack. x.com/FT/status/2097…https://x.com/ai/status/20993808232294033080.00306952026-09-14T14:51:20.812988+00:002x.comlawfirmnvidiaai600https://pbs.twimg.com/profile_images/1831182575689240576/7HmOAeIK_mini.jpg2026-09-14T14:51:21.015045+00:00695@ayushtweetshereAyushthis is Dario and sama's worst nightmare - A law firm buying Nvidia servers. Latham & Watkins is building an in-house AI stack.. this is US’s second-largest law firm with $8.3 Billion in revenue last year And now it has - - Nvidia hardware it controls - open-weight modelshttps://x.com/ayushtweetshere/status/20993808232294033080.00306962026-09-14T14:51:20.604386+00:002x.comnvidialawfirmai693@aianand iyerLatham & Watkins, the 2nd largest US law firm, is buying Nvidia hardware to fine-tune open weights in-house. Open weights + proprietary data + local compute = enterprise sovereign AI stack.https://x.com/ai/status/20992497818268017600.00306942026-09-14T14:51:05.152932+00:002x.comlawnvidiaai598https://pbs.twimg.com/profile_images/1831182575689240576/7HmOAeIK_normal.jpg2026-09-14T14:51:05.371173+00:00692@FTFinancial TimesLatham & Watkins buys Nvidia servers to set up in-house AI systemshttps://x.com/FT/status/20979931196522663020.00402026-09-14T14:50:48.316603+00:002x.comlatham--watkinsnvidiaai666@ayushtweetshereAyush<p>This is what Dario and Sama fear - - the cost of fine tuned open source models trained on custom data was 95% less than the frontier models - And they performed better than the frontier models - And they could train them in less than 48 hours The big AI labs have no moat.. Even at the enterprise level.. Any sane company would prefer a custom trained open source model instead of paying API pricing to Anthropic and Open AI If open source keeps growing at the pace it is, frontier labs' business collapses.. only way out is oldest trick in manipulation - FUD -&gt; Fear, Uncertainty, Death Step 1 - Seed the psyop - Test a swarm of agents trained to "hack" - The swarm does what its supposed to do - hack - Publish a report - AI is too dangerous Step 2 - weed out the doubters - Get an employee to quit your company - Go viral saying both companies are not doing enough to self regulate AI - AI is dangerous - will kill humanity by the end of the decade Step 3 - Prepare for fruition - write an "essay" to "pace the frontier" - Advocate regulation - Only a few holier than though companies get to build AI - Scare the world into agreeing with AI Doom - Effectively turn your competitors into criminals (genius business strategy) Endgame - - Become a cartel that controls AI (and the world) - Raise prices (coz business is unsustainable... ofc) - IPO (offload liability of said unsustainable business to the public) Laugh all the way to the bank Win!</p>https://x.com/ayushtweetshere/status/20990000910593027810.00602026-09-14T09:04:56.854564+00:002x.comaiopen-sourcemodels582https://abs.twimg.com/emoji/v2/svg/1f911.svgMoney-mouth face2026-09-14T09:04:57.305388+00:00664@L3HarrisTechL3HarrisTech<p>Palantir says it helped them beat frontier AI models in less than 2 days, at 95% lower cost: "When we fine-tuned open source models trained on our own data, we were able to outperform the frontier models in less than 48 hours." "The cost of our fine-tuned open source model was 95% lower than the frontier models we were using." "AI is a commodity. It's all about the data. We view our data as a corporate asset. It's our unique hard-earned knowledge." "We believe American defense companies should not be a vassal for frontier AI labs, handing over our data and institutional knowledge, hoping to rent back the intelligence it creates." " We own the model, we own the compute, we own the advantage."</p>https://x.com/L3HarrisTech/status/20981246652607366840.00602026-09-14T08:55:34.112713+00:002x.compalantirdefenseai579IMAGE/media/content/664/image.pngimage.pngpng2026-09-14T08:57:01.574726+00:00637@OrcaRouterOrcaRouterWhy does everyone suddenly want to pace AI? Recurrent Looped Transformers. We now have Transformers with potentially unbounded reasoning depth: same weights, more loops, more compute - toward Super Intelligence. The scary part isn’t intelligence scaling. It’s alignment x.com/orcarouter/sta…https://x.com/OrcaRouter/status/20992242601371161310.00456362026-09-14T07:20:43.060437+00:002x.comalignmentaitransformers562https://pbs.twimg.com/profile_images/2054951018102964224/GLp-Nttr_mini.jpgSquare profile picture2026-09-14T07:20:43.356941+00:00636@OrcaRouterOrcaRouterRumor is next-gen models at OpenAI + Anthropic are showing emergent misalignment beyond what’s been publicly reported. Technical backgrounds sound particularly interesting: → Misaligned behavior appearing in CoT → Existing monitors reportedly have insufficient AUROC → Exploit-heavy RL may be making this generation worse → Directly training against “bad” CoT risks teaching models to hide intent → Proposed path: sublingual monitoring — latent/J-space probes instead of trusting visible reasoning Apparently both labs are reluctant to optimize the CoT itself (could make the model conceal problematic reasoning in the future). Alignment just became a representation-space problem.https://x.com/OrcaRouter/status/20992242601371161310.00456372026-09-14T07:20:42.323367+00:002x.comalignmentaitech626https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA<p>JUST IN: Dario Amodei, Sam Altman, and Elon Musk have all agreed that the AI industry should “slow down”Source: TrendSpider@TrendSpider</p>https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00502026-09-13T13:23:56.219621+00:001linkedin.comslowdownaitrendspider552IMAGE/media/content/626/image.pngimage.pngpng2026-09-13T13:24:27.348943+00:00623@yifanzhang_Yifan ZhangWe are at the dawn of Superintelligence. Introducing the Recurrent Looped Transformer (RLT), We now have Transformers with Infinite Reasoning depth. From now on, we should pace progress at the Open Frontier of Superintelligence, Until Safe Superintelligence is achieved. https://github.com/yifanzhang-pro/recurrent-looped-tranformer…https://x.com/yifanzhang_/status/20988862680339456100.00402026-09-13T08:50:04.752887+00:002x.comsuperintelligenceaitransformer549https://pbs.twimg.com/media/HSC91w8bwAELtrN?format=jpg&name=smallImage2026-09-13T08:50:06.653411+00:00620@GoogleResearchGoogle ResearchCan AI build new knowledge from experience? Join Google researchers Dima Damen and Viorica Patraucean alongside Yilun Du for the Panel today at 10:00am CEST in the Arena Room as they discuss the future of AI learning loops: https://eccv.ecva.net/virtual/2026/panel/6160 … @GoogleDeepMindhttps://x.com/GoogleResearch/status/20986528291875350960.00502026-09-12T12:55:19.433014+00:002x.comresearcheccv2026ai619@NeoCognitionNeoCognitionIntroducing ApprenticeBench: computer use + continual learning on a real job. We show Fable 5.1 and GPT-6 Astra can now continually learn on a job and surpass human professionals. A decisive step change in AI's job readiness. No FDEs. Agents deploy themselves into the job.https://x.com/NeoCognition/status/20981471856575653910.00502026-09-12T12:54:09.926795+00:002x.comlearningjob-readinessai547https://pbs.twimg.com/media/HR4cclKacAABQJV?format=jpg&name=smallImage2026-09-12T12:54:10.172230+00:00611https://www.linkedin.com/in/florianbansac/Florian BansacClaude just opened a Northrop Grumman position in a $100,000 AI trading experiment that is up 26%.The experiment started in late November and has beaten the S&P 500 so far. The new trade is only a starter position, which makes the decision easier to assess over time.This is a stronger agent test than polished market analysis. Claude must allocate limited capital, accept visible risk, and produce a measurable result.AI agents become credible when their decisions have a cost and a benchmark.https://www.linkedin.com/feed/0.00452026-09-12T12:43:15.450733+00:001linkedin.comexperimentaitrading544https://media.licdn.com/dms/image/v2/D4E22AQFGw4SAw08K9A/feedshare-shrink_800/B4EaCRXBv3HQAc-/0/1789145128182?e=1790812800&v=beta&t=mOBt9XU7CC78h1yFw5bPg78Kmi7J00wHwMEgneFifCkView image2026-09-12T12:43:16.552410+00:00605@ControlAIControlAI70+ cross-party MPs and peers are calling on the UK Prime Minister to lead an international agreement prohibiting the development of superintelligent AI, while preserving strategic AI ambitions. They urge him to adopt the Artificial Superintelligence Security Bill. Threadhttps://x.com/ControlAI/status/20984595249065863510.00502026-09-12T08:38:07.804789+00:002x.comuksuperintelligenceai538https://pbs.twimg.com/media/HR85H9laIAAX7Pr?format=jpg&name=smallImage2026-09-12T08:38:08.013232+00:00525https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterThe AI Revolution is fueling a record financing boom in the US:US high-grade corporate bond issuance is expected to surge to ~$215 billion in September, the largest September issuance on record.This is set to surpass the previous September high of ~$205 billion, posted in 2025.The current estimate is also more than triple the ~$70 billion issued in September 2022.This follows $145.2 billion of US high-grade bonds issued in August, the highest August total on record, as AI-related data center buildouts continue to fuel corporate borrowing.US companies have already borrowed more than $410 billion so far in 2026, specifically to fund data centers and other AI-related investments.Meanwhile, retail investors have purchased more investment-grade bonds this year than in any full year in data going back to 2010.The AI boom is driving a record corporate borrowing spree.https://www.linkedin.com/company/the-kobeissi-letter/0.00602026-09-09T01:08:50.075239+00:001linkedin.comcorporate-bondsaifinance471IMAGE/media/content/525/image.pngimage.pngpng2026-09-09T01:09:43.573724+00:00521https://www.linkedin.com/in/wei-li-a93561b/Wei Li<p>AI is starting to show up beyond the AI trade.The supply chain remains in a league of its own. But more recently, earnings at S&amp;P 500 companies adopting AI have begun to pull away from non-adopters (chart).Commoditisation risk doesn’t equal destruction of economics. It means a potential transfer of rents.</p>https://www.linkedin.com/feed/0.00452026-09-08T17:42:21.604433+00:001linkedin.comsupply-chaineconomicsai468IMAGE/media/content/521/image.pngimage.pngpng2026-09-08T17:43:08.512860+00:00515https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA🚨 AI token demand could reach 4,000 quadrillion tokens annually by 2030.That is Evercore’s base case—more than 20× today’s consumption. Goldman Sachs projects a similar 24× increase.The biggest driver? Agentic AI.AI agents could become the largest source of token demand by 2030. They operate continuously—and a single agentic task can consume 10–100× more tokens than a chatbot response.But the forecast excludes one potentially enormous category: physical AI—robots, autonomous machines and intelligent industrial systems.Every token ultimately requires physical infrastructure:🔹 GPUs🔹 Memory🔹 Networking🔹 Electricity🔹 Data centresThat infrastructure largely does not exist yet.AI-related stocks will experience corrections. But forecasts for underlying demand continue to be revised higher—not lower.The AI infrastructure cycle may have much further to run than markets currently assume.Source: Kyle Reidhead | Milk Road@KyleReidhead·https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00502026-09-08T12:46:11.032660+00:001linkedin.cominfrastructuretoken-demandai462IMAGE/media/content/515/image.pngimage.pngpng2026-09-08T12:46:37.251229+00:00511@unknownUnknownTrump Faces Fresh Calls for Removal Over AI-Fueled Posting Frenzyhttps://x.com/thedailybeast/status/20969566628966197570.00455102026-09-08T09:21:56.349269+00:002x.comtrumpremovalai459https://pbs.twimg.com/card_img/2096956663890591746/mbs6qy3O?format=jpg&name=smallU.S. President Donald Trump speaks during an event after signing an executive order renaming Lake Ontario as Lake America in the Oval Office at the White House in Washington, D.C., U.S., August 27, 2026.2026-09-08T09:21:56.618837+00:00510@thedailybeastThe Daily BeastThe 80-year-old president is facing calls to be removed from office via the 25th Amendment after he spent an entire day posting deranged AI-generated slop on Truth Social.https://x.com/thedailybeast/status/20969566628966197570.00455112026-09-08T09:21:56.004115+00:002x.compresidenttruth-socialai503@McKinseyMcKinsey & CompanyAI is forcing software leaders to rethink the team itself—not just how work gets done. The next advantage may come from redesigning roles, ownership, and team structures around what people and agents do best. https://mck.co/4zT8qOMhttps://x.com/McKinsey/status/20969318123290178260.00502026-09-08T09:02:59.336506+00:002x.comsoftwareaiteam453https://pbs.twimg.com/media/HRnMXy8XIAI_5nY?format=png&name=smallBar chart showing developers and product managers reallocating work time from execution to strategic tasks due to AI, with top accelerators shifting more hours.2026-09-08T09:03:01.525783+00:00494@McKinseyMcKinsey & CompanyIt’s easy to focus on what an AI agent can do. The harder question is whether the economics hold up at scale. Leaders need to look beyond model costs and rethink how work is allocated between people and AI to find where agentic workflows can create value. https://mck.co/4qQuTbbhttps://x.com/McKinsey/status/20970523666545462780.00502026-09-08T05:22:21.790229+00:002x.comeconomicsaiworkflows444https://pbs.twimg.com/media/HRo6A-MXMAE6seu?format=png&name=smallBar chart showing annual run cost per agent in banking, highlighting human oversight as a major variable cost between 70-75%.2026-09-08T05:22:22.101907+00:00486https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIAAI Agents are using 1400% more tokens today than February, 500% more than human.Remember, this is inference token traffic. Every token lands in the KV cache which is a memory problem.Will supply be able to catch up?Source: Trade Whisperer@TradexWhisperer$MU$SKHY$DRAMhttps://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/0.00352026-09-08T05:11:17.782465+00:001linkedin.comtokensaiinference437IMAGE/media/content/486/image.pngimage.pngpng2026-09-08T05:11:48.000025+00:00478@DeryaTR_Derya Unutmaz, MD<p>This is a major milestone in the age of AI and toward our ultimate goal of completely reversing aging! developed the first AI drug, Rentosertib, now in a Phase 3 clinical trial, and it just showed something remarkable: a reversal of biological age in people! Six different aging clocks, developed by six independent groups, all showed reversal after just 12 weeks of treatment, with participants measuring 3–4 years younger biologically! Congratulations to my friend Alex and the entire @InSilicoMeds team on this landmark achievement!</p>https://x.com/DeryaTR_/status/20969890190746297830.00404792026-09-08T04:42:34.232787+00:002x.comagingmilestoneai463@FedeeForteFederico Forte<p>Calculations of AI's impact on employment are being refined. I came across this little map that quantifies its effect on each occupation in the US: https://liminal-capital.com/apps/ai-job-map.html It breaks down its impact into 3 distinct effects: 1) Substitution effect (S): AI replaces human tasks. This would destroy jobs. 2) Complementary effect (C): AI can perform human tasks but needs the presence of a human to carry them out. This has the potential to create jobs or at least not make them disappear. 3) "Isolation" of AI: there's no way yet to automate that job. For example, the "financial advisor" role is assigned 48% substitution (S) and 52% complement (C); for "educators": S: 29% / C: 54%; for web developers S: 94% / C: 3%, and so on. In this paper, the authors explain how everything is calculated: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7195359 Two more reflections from there that I found interesting: 1) what AI does/doesn't do are "tasks," not "jobs." So it's important to distinguish which part of each job is replaceable/enhanceable/isolated from AI in order to quantify its overall automation risk. 2) AI's empirical impact is seen more in the new income of employees and not so much in people already employed, so a regression on new hires can/will yield different results than the conclusions that can be drawn from calculations on the total mass of employees. I think we still don't have all the info available here to replicate something like this, but it would be interesting.</p>https://www.liminal-capital.com/apps/ai-job-map.html0.00452026-09-07T04:49:04.770403+00:002x.comemploymentjobsai421IMAGE/media/content/463/image.pngimage.pngpng2026-09-07T04:50:57.917022+00:00458@IntCyberDigestInternational Cyber DigestThe creator of the C++ programming language, Bjarne Stroustrup, says AI-generated code is not successful and calls the idea of using natural language as a programming language idiotic. He says humans will still write code and use abstraction. According to him, AI generates bloated code with more bugs and security holes, making it hard to validate. He also says the senior developers needed to validate it are starting to retire because they don’t want to deal with validating something that changes every time you change your prompt. He goes on to say that AI is not good at writing safety-critical or performance-critical code: “Now, let’s say that 70 or 80% of the world’s code doesn’t fit that pattern, but it’s that 10, 20% of the code that I’m interested in.”https://x.com/IntCyberDigest/status/20966541495006987610.00502026-09-07T04:22:36.556091+00:002x.comaiprogrammingc++416https://pbs.twimg.com/media/HRjPsgxbYAAtksT?format=jpg&name=smallImage2026-09-07T04:22:36.954435+00:00444@thesupermanmxSupermanChina published the most uncomfortable paper on vibe coding. ETH Zurich tested 100 developers in a controlled, commercial-grade vibe coding environment to see who actually succeeds. The findings are brutal. The researchers tracked computer science achievement, written communication skills, and general cognitive reasoning. They wanted to see what actually predicts vibe coding proficiency when you never touch a line of source code yourself. Two major predictors emerged. Written communication proficiency mattered. The ability to structure thoughts and articulate intent unambiguously in text directly impacts what the AI builds. But that wasn't even the main takeaway. Computer science achievement was a massive, dominant predictor of success. Even when researchers controlled for general intelligence and reasoning skills, CS background still heavily dictated who built working software and who completely crashed. In fact, CS knowledge contributed roughly twice the unique predictive variance of writing skills alone. Why? Because vibe coding isn't about writing code. It’s about debugging logic. When an AI agent builds a complex application and quietly breaks an edge case under the hood, a non-technical user looks at the glowing UI and assumes it works. They don't know what questions to ask. They don't know what logic to challenge. They lack the mental models to recognize architectural catastrophe. You can prompt your way past syntax. You cannot prompt your way past a fundamental lack of engineering intuition. The hype told us that learning to code is dead because language is all you need. The data just proved the opposite. To truly master the vibe, you still need to understand how the machine thinks.https://x.com/thesupermanmx/status/20961684388300802780.00702026-09-06T14:15:15.411986+00:002x.comaivibe-codingcomputer-science408https://pbs.twimg.com/media/HRcWDk8bgAAkVOB?format=jpg&name=smallImage2026-09-06T14:15:15.733818+00:00439https://www.linkedin.com/in/sheikh-marzan/Marjanul Islam<p>🔴 OpenAI GPT-6 Astra was trained on 100,000 GPUs and cost $1 billion, yet it is only 3% better than an open-source model.In other words, an open-source model can achieve 97% of GPT-6 Astra’s capabilities. So, does training a new model by investing billions of dollars really make any sense?To get the answer, you first have to realize that continuous improvement in AI models requires more data, more GPUs, and longer training.An open-source model means the parameters—the rules of the AI game—are open. You can see them and inspect them. With a closed model, those parameters are not disclosed to the public.The fact is, to release the parameters of an open-source model, the developers also need to train the model. Otherwise, how would they obtain those parameters?And it is highly likely that almost all open-source models are, in some form, derived from closed models—and that is completely fine.So, without continuous improvement in frontier closed models, there will also be less and less improvement in open-source models. Without additional training, development eventually comes from algorithmic efficiency gains, and efficiency gains cannot increase forever.So yes, the world needs closed models that are vastly more powerful than GPT-6 Astra.And a 3% improvement in intelligence is huge—enormous.How?The difference between your local zoo and your community is only a few percentage points of intelligence.A tiny change or improvement in intelligence can create a sea of difference.So yes, it is totally fine to invest billions of dollars for a 3% improvement.GPT-6 Astra cost $1 billion to train. Soon, we will see $10 billion spent to train a single model.And by 2030, we will see a model that requires $100 billion just to train.AI isn’t stopping. No way.</p>https://www.linkedin.com/in/sheikh-marzan/0.00602026-09-06T07:37:13.438357+00:001linkedin.comaiopenaigpt-6405https://media.licdn.com/dms/image/v2/D5622AQGlIpVy6GecQg/feedshare-shrink_800/B56aByRGNtIMAc-/0/1788623479922?e=1790208000&v=beta&t=KiP_Qaot-iQMJ-jeB2Bbaj08hajC2oHgO7-Rju_zI6AView image2026-09-06T07:37:13.761380+00:00434https://www.linkedin.com/company/the-kobeissi-letter/The Kobeissi LetterAI adoption is spreading far beyond tech:80.6% of US technology and media businesses now have paid AI subscriptions, the highest percentage on record.Finance and insurance companies follow at 73.1%, up from ~60.0% in December 2025.Meanwhile, 60.4% of US manufacturing firms are paying for AI subscriptions, their highest proportion on record and the 3rd-highest among major sectors.This percentage has nearly doubled since the start of 2025.Over the same period, the proportion of retail companies paying for AI subscriptions has risen +20 percentage points, to a record 49%.Non-tech firms are now quickly adopting AI.https://www.linkedin.com/feed/0.00602026-09-06T07:09:36.938880+00:001linkedin.comtechnologyaibusiness433https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA⚠️ The bond market is starting to put a price on the AI boom.S&P is assessing how much investment capacity hyperscalers have before their credit ratings come under pressure.Alphabet and Microsoft retain enormous headroom. Meta still has a reasonable cushion. Amazon has far less room for error.But Oracle is the real pressure point.Its adjusted debt-to-EBITDA could approach 4.5x—the level associated with pressure on its BBB- rating, the final rung of investment grade.A downgrade could trigger forced selling, wider spreads and higher funding costs—precisely when Oracle must keep investing heavily in AI and cloud infrastructure.The broader risk is timing: capex, leases and power contracts are locked in today, while AI revenues may arrive years later.With hyperscalers potentially spending more than $7 trillion over five years, unlimited AI financing may be coming to an end.The bond market could impose discipline first.Source: FThttps://www.linkedin.com/feed/0.00702026-09-06T07:09:12.425190+00:001linkedin.cominvestmentaifinance401https://media.licdn.com/dms/image/v2/D4E22AQEo_uy8Ne3sdw/feedshare-shrink_160/B4EaByhvxAK4Ak-/0/1788627844466?e=1790208000&v=beta&t=8jfFmJ_rI3ulgQLwKV2m5B11OYM_-pct-jxSisWHhVYView image2026-09-06T07:09:14.337192+00:00423https://www.linkedin.com/in/siegfried-spitzer-0087629/Siegfried Spitzer<p>So why hasn't anyone launched the obvious Inverse Burry Tracker?Introducing the Cassandra Unchained Inverse Tracker ETF (ticker: LMAO). Investment objective: results approximately opposite to those of a man who has been calling the top since January 2023.The index sleeve: he tweeted "Sell" in 2023, the S&amp;P had one of its best years of the decade. He bought $1.6B notional in SPY/QQQ puts, then explained they were hedges. This August he rolled his QQQ puts out to June 2027 with the Nasdaq up 19% YTD. LMAO will collect the rent.The AI sleeve: 80% of the portfolio in Nvidia and Palantir puts. Nvidia then reported an 85% revenue surge. Micron reported record revenue. His response: add to the Micron short at ~$934 while the stock had already tripled this year, against 40 of 45 analysts rating it a Buy. Palantir, per him, is worth a fraction of its price long term.The long sleeve: DraftKings and Flutter, on the theory that regulators will move quickly. We short those too. Diversification. #Lululemon is down 58.8% since he opened his long in Q2 2025. It is now his largest position at 17.4% of the portfolio. After Thursday's earnings sent it to an eight-year low, he announced he'd buy more below $100. LMAO will be there to sell it to him.And unlike the Cramer fund, the data feed is free. Since deregistering Scion in November 2025, he publishes his trades on Substack, sometimes with entry prices.Risk factor: he's early rather than wrong.</p>https://lnkd.in/p/d4Fdvx320.00802026-09-06T05:14:11.236473+00:001linkedin.comailululemonetfmarket394IMAGE/media/content/423/image.pngimage.pngpng2026-09-06T05:15:25.212451+00:00415https://www.linkedin.com/in/charles-henry-monchau-cfa-cmt-caia-4003096/Charles-Henry Monchau, CFA, CMT, CAIA<p>⚠️ THE AI BOOM IS MORE CONCENTRATED THAN INVESTORS REALIZE.Just 1% of enterprise customers generate roughly 80% of OpenAI and Anthropic’s combined enterprise revenue, according to Ramp AI Index data.This concentration has persisted for two years—and is unmatched across other software categories tracked by Ramp.The biggest spenders are primarily technology and AI-related companies.That creates a major vulnerability ahead of potential OpenAI and Anthropic IPOs.If markets correct and even a handful of top customers cut spending, AI revenues could slow sharply.When growth relies on so few customers, the unwind can be just as concentrated.Source: Global Markets Investor, Ramp</p>https://lnkd.in/p/dmSDkYf50.00802026-09-05T14:13:31.573325+00:001linkedin.commarketinvestmentai386https://media.licdn.com/dms/image/v2/D4E22AQFhTeKm8ZnRdA/feedshare-shrink_480/B4EaBukkqgH8Ag-/0/1788561476542?e=1790208000&v=beta&t=IXd4Ms_bttZ6bP353PVF5hpn3_qJPYShgsH3sGpVLT4View image2026-09-05T14:13:31.927007+00:00395@Gartner_incGartner<p>What if AI isn't reducing workforce costs, but simply shifting them? Gartner predicts that up to 30% of roles displaced by AI will be rehired by 2029, often at a premium. Treating AI as a cost-cutting tool alone can undermine ROI; optimizing workforce costs is what drives</p>https://x.com/Gartner_inc/status/20959148142330188370.00502026-09-05T10:37:21.139332+00:002x.comaiworkforcecosts365IMAGE/media/content/395/image.pngimage.pngpng2026-09-05T10:37:43.229856+00:00393@thesupermanmxSuperman<p>MIT mathematically proved that AI will destroy the democracy. A Nobel-winning MIT economist published a terrifying paper called "Automation and Repression" As AI and automation replace human labor, wealth concentrates heavily in the hands of a tiny group of capital owners. Inequality skyrockets. When inequality hits a critical mass, workers realize they are being crushed and the threat of a popular revolt spikes. Faced with that threat, the ruling elite are forced to make a choice. They can redistribute the wealth through taxes, or they can use force to keep people down. The paper mathematically proves a dark reality: there is a direct, inescapable link between automation and political repression. The more capital accumulates through AI, the more the elite prefer repression over redistribution. Why? Because sharing the wealth cuts into their power. Funding a police state protects it. It gets worse. The authors modeled what happens when an economy starts inside a clean, stable democracy. As automation advances and capital concentrates at the top, the math shows that the elite eventually find democracy to be a liability. They stop supporting democratic systems. They back a coup. They install a repressive system just to protect their automated wealth. Nobody is voting to end democracy. The technology’s economic incentives just make authoritarian control the logical next step for survival.</p>https://x.com/thesupermanmx/status/20959173717353144560.00602026-09-05T09:04:33.871980+00:002x.commitairepression363https://pbs.twimg.com/media/HRYxsm1bgAAOEUW?format=png&name=smallImage2026-09-05T09:04:35.406485+00:00380@aliansarinikAli Ansariwe’re hiring 10,000 robotics trainers in the next 7 days. $50–$90/hour, accepting applicants globally. you’ll review and label videos of robots performing tasks to help them improve. no prior AI experience required. an entirely new category of work is emerging around teaching robots how to interact with the world. application link in the comments below.https://x.com/aliansarinik/status/20960329081969095390.00502026-09-05T05:22:59.418021+00:002x.comhiringairobotics367@GlobalMktObservGlobal Markets Investor🔴Nvidia is trailing the very AI boom it's supposed to be powering: Nvidia, $NVDA, shares are up just ~20% over the last 12 months, the weakest performer among a group of AI-linked indices tracked by Bloomberg. TAP IMAGE TO SEE FULL INSIGHT👇https://x.com/GlobalMktObserv/status/20951494628175257770.00702026-09-04T12:30:02.703080+00:002x.commarketsnvidiaai339https://pbs.twimg.com/card_img/2093744365021667328/FwxHhNrd?format=png&name=small2026-09-04T12:30:03.859870+00:00366@GlobalMktObservGlobal Markets Investor⚠️AI pricing is collapsing, raising questions about AI returns: The LLM Token Expenditure Index, which tracks the market price companies pay for AI model output, has fallen to just $0.97, its lowest level since the index was created late last year and more than -50% below its summer peak. Token prices are collapsing as cheaper models, open-source Chinese competitors and falling inference costs make AI increasingly commoditized. That is great for AI users, but potentially much worse for model providers, as falling prices directly erode revenue and pricing power while massive compute commitments remain largely fixed. The bigger concern is what this means for the returns on the hundreds of billions being poured into AI infrastructure, if the price companies can charge for AI output keeps falling. The AI boom depends not just on demand, but on that demand generating enough revenue to justify the enormous capital spending behind it. If AI output becomes a commodity faster than demand can absorb the falling prices, the biggest AI valuations could face a serious reality check.https://x.com/GlobalMktObserv/status/20951406898770780560.00502026-09-04T12:29:34.530117+00:002x.compricinginfrastructureai338https://pbs.twimg.com/media/HRNvWs5WQAISgsL?format=jpg&name=smallImage2026-09-04T12:29:34.825391+00:00341@fcholletFrançois CholletGPT-6 Astra represents a step-function change in model capability for interactive reasoning problems. It scores 66% on ARC-AGI-3 using our standard harness, and nearly 100% with a continuous conversation harness and custom compaction, at a cost of roughly $360 per game. In fact,https://x.com/fchollet/status/20955984511156143710.00353422026-09-04T06:25:04.663372+00:002x.comarc-agi-3aigpt-6317https://www.linkedin.com/in/edzitron/Ed ZitronHere's my segment from earlier in the year with Bloomberg, covering how Anthropic and OpenAI are dangerous and unsustainable companies that shouldn’t IPO. The AI bubble is a con and retail investors are the marks. AI doesn’t have ROI, it’s nothing like AWS/Uber, and it’s got no post-bubble recovery story.https://www.linkedin.com/feed0.00802026-09-04T05:31:27.753406+00:001linkedin.comipoaibloomberg292https://media.licdn.com/dms/image/v2/D5605AQEAvyojraXZ3g/videocover-low/B56aBjlnRBGUBI-/0/1788377206625?e=1789106400&v=beta&t=6XOJn05p7mOr_aAinB0yv8fT2_NJjPfDvhLDYooThLY2026-09-04T05:31:28.025196+00:00308@GoogleAIGoogle AIIntroducing WeatherNext 3️⃣— our most advanced global weather AI model yet from @GoogleDeepmind and @GoogleResearch With prediction capabilities that are up to 5x sharper than WeatherNext 2, the model generates a forecast with high spatial resolution in order to catchhttps://x.com/GoogleAI/status/20955449441907880640.00502026-09-04T05:19:57.674675+00:002x.comweatheraiforecast286https://pbs.twimg.com/media/HRTe9Epa0AA5vOP?format=jpg&name=smallImage2026-09-04T05:19:58.161199+00:00306https://www.linkedin.com/company/the-kobeissi-letter/posts/The Kobeissi Letter<p>AI and software stocks are now rising together.The 1-month correlation between AI-related stocks and software stocks has risen +0.70 over the last month, to +0.15, marking its largest monthly increase since July 2025.By comparison, the 1-month correlation was as low as -0.56 in July 2026, as investors increasingly viewed AI as a threat to traditional software businesses.As a result, hedge fund exposure to software and services stocks declined -5 percentage points over the 12 months ending July, to just ~1% of total global hedge fund market exposure, near its lowest level on record.The recent increase in correlation comes as some software firms previously viewed as vulnerable to AI are actually finding ways to use the technology to strengthen their existing businesses, improve productivity, and defend their competitive advantages.Meanwhile, the US software ETF, IGV, is up +39% since its April low, recovering most of its drawdown that began in October 2025.The AI trade may be shifting from disruption to adaptation.</p>https://lnkd.in/p/d7MSRunu0.00602026-09-04T05:14:38.198754+00:001linkedin.comsoftwarestocksai284https://media.licdn.com/dms/image/v2/D4E22AQH0oAkq6DJDnw/feedshare-shrink_800/B4EaBpge6VJwAg-/0/1788476518308?e=1790208000&v=beta&t=3U0mH0XbZVc70D8gv3-C7sWbDSSQ0NKMfvwdPOG0EiQView image2026-09-04T05:14:38.456677+00:00303https://www.linkedin.com/in/akharazian/Ara KharazianNew from Ramp data: the latest threat to the AI trade. AI companies' revenues are heavily dependent on a small set of customers. 80% of OpenAI and Anthropic's enterprise revenues come from 1% of their customers, and it's not getting better.This is a level of concentration risk unseen in any other software category we track. The companies in the top 1% skew heavily toward the tech sector and AI products and services. What happens in a market correction? All these companies are highly correlated, and an increasing share of our economy is invested in them. Especially as we approach blockbuster IPOs for OpenAI and Anthropic.The latest from Ramp AI Index: ramp.com/data/ai-indexhttps://www.linkedin.com/feed/0.00802026-09-03T20:37:08.445836+00:001linkedin.comairampeconomy281https://media.licdn.com/dms/image/v2/D4E22AQFrl191-lor7Q/feedshare-shrink_800/B4EaBjMnLmHoAc-/0/1788370646071?e=1790208000&v=beta&t=pSenuI-tY-FaLjt3x4ndh9yeaoHDTAFFnfGWhlN4uBQView image2026-09-03T20:37:10.208310+00:00298@NewsFromGoogleNews from GoogleIntroducing WeatherNext 3, our most advanced and accurate global weather AI model to date, from @GoogleDeepMind and @GoogleResearch. This new forecasting model learns directly from real-time observations, and uses raw satellite data to produce a forecast every hour in high resolution. These breakthroughs mean it can provide timely and more localized predictions, bringing even more reliable forecasts across Google products worldwide, including Search, Gemini and Maps.https://x.com/NewsFromGoogle/status/20955303093997572010.00702026-09-03T17:17:56.792355+00:002x.comweatheraiforecast273IMAGE/media/content/298/image.pngimage.pngpng2026-09-03T17:19:00.940202+00:00297@jeffreyleefunkjeffrey lee funkLLM Token Prices collapsed by almost 60% in Q2, and will likely force a cut in AI capex. Global interest rates are soaring, which means higher costs for hyperscalers. The AI bubble burst is likely already in progress, will likely accelerate in Q3 and Q4.https://x.com/jeffreyleefunk/status/20953315678247118790.00502026-09-03T17:16:51.510007+00:002x.comtechnologyaifinance272https://pbs.twimg.com/media/HRQcaOfaIAAnBlx?format=jpg&name=smallImage2026-09-03T17:16:52.216594+00:00286@arakharazianAra KharazianNew from Ramp data: the latest threat to the AI trade. AI companies' revenues are heavily dependent on a small set of customers. 80% of OpenAI and Anthropic's enterprise revenues come from 1% of their customers, and it's not getting better. This is a level of concentration risk unseen in any other software category we track. The companies in the top 1% skew heavily toward the tech sector and AI products and services. What happens in a market correction? All these companies are highly correlated, and an increasing share of our economy is invested in them. Especially as we approach blockbuster IPOs for OpenAI and Anthropic.https://x.com/arakharazian/status/20952044526091715550.00502026-09-03T04:48:39.024372+00:002x.comriskaifinance261https://pbs.twimg.com/media/HROpDFfXgAE_M1T?format=jpg&name=smallImage2026-09-03T04:48:39.341728+00:00279@OpenAIOpenAIAs we prepare to release Astra, we’re focused on making increasingly capable AI safe and broadly accessible. Astra represents a significant advance in cybersecurity capability, reaching the Critical threshold under our Preparedness Framework. We're previewing how we evaluatedhttps://x.com/home0.00402782026-09-02T13:35:10.508882+00:002x.comcybersecurityaiastra257https://pbs.twimg.com/profile_images/1885410181409820672/ztsaR0JW_mini.jpgSquare profile picture2026-09-02T13:35:10.678968+00:00278@levieAaron LevieAI for cyber is about to go vertical. The models increasingly becoming insanely good at finding and exploiting vulnerabilities. Frontier models are ahead, but we’re already seeing that open weights is not far behind. Most enterprises are already inundated with cyber discoveries,https://x.com/levie/status/20950246994411196120.00402792026-09-02T13:35:10.224276+00:002x.comcybersecurityaivulnerabilities256https://pbs.twimg.com/media/HRMF3JGa4AArAwx?format=jpg&name=smallImage2026-09-02T13:35:10.485221+00:00269@MultiverseCompuMultiverse Computing🚀 Introducing Quasar 438B, the top European AI model. Externally validated by @ArtificialAnlys: in their latest benchmark comparison, Quasar is the strongest-performing European model evaluated. It is the first large model released by Multiverse Computing. 438 billion parameters, English and Spanish, built for enterprise-scale agents, coding and complex multi-step workloads.https://x.com/MultiverseCompu/status/20950610115018466310.00702026-09-02T12:51:21.833832+00:002x.comeuropeanbenchmarkai242https://abs.twimg.com/emoji/v2/svg/1f680.svg🚀2026-09-02T12:51:22.092663+00:00243https://pbs.twimg.com/media/HRMm2Y6WwAEkhQF?format=jpg&name=smallImage2026-09-02T12:51:22.127519+00:00265@miclchenMichael L. ChenAll three pillars of a safety case look about to fall. We are rather likely to have highly capable, poorly monitorable, dubiously aligned AI agents working autonomously inside the world's most consequential organizations.https://x.com/miclchen/status/20949976712326105290.00402026-09-02T12:48:27.343249+00:002x.comsafetyalignmentai213@productaizeryPatrick SentiExpect AI models to be unaccessible in the EU soon.https://x.com/productaizery/status/20939783414978805790.00402142026-08-31T08:00:47.057664+00:002x.comeuaimodels194https://www.linkedin.com/in/manuelheinkel/Manuel HeinkelA site with 1.62M Google impressions just dropped to almost zero.Every post was 100% AI-generated.And they are not the only ones. Over the past few days, several website owners have reported similar drops.The decline came after Google’s spam update 3 days ago.⚠️ 𝗧𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝗻𝗼𝘁 𝗺𝗲𝗮𝗻 𝗚𝗼𝗼𝗴𝗹𝗲 𝗶𝘀 𝗽𝗲𝗻𝗮𝗹𝗶𝘇𝗶𝗻𝗴 𝗔𝗜 𝗰𝗼𝗻𝘁𝗲𝗻𝘁 𝗶𝗻 𝗴𝗲𝗻𝗲𝗿𝗮𝗹. 𝗜𝘁 𝗶𝘀 𝘁𝗮𝗿𝗴𝗲𝘁𝗶𝗻𝗴 𝘀𝗰𝗮𝗹𝗲𝗱, 𝘂𝗻𝗼𝗿𝗶𝗴𝗶𝗻𝗮𝗹 𝗰𝗼𝗻𝘁𝗲𝗻𝘁 𝘁𝗵𝗮𝘁 𝗮𝗱𝗱𝘀 𝗹𝗶𝘁𝘁𝗹𝗲 𝘃𝗮𝗹𝘂𝗲, 𝘄𝗵𝗲𝘁𝗵𝗲𝗿 𝗶𝘁 𝘄𝗮𝘀 𝗰𝗿𝗲𝗮𝘁𝗲𝗱 𝘄𝗶𝘁𝗵 𝗔𝗜 𝗼𝗿 𝗻𝗼𝘁.At 🥥cocoduro, we use a custom AI skill for our blog posts. It is a reusable set of instructions that tells AI how to perform a repetitive task consistently.It helps us:• Interlink existing posts• Structure content clearly• Improve titles and technical SEO• Work faster But we still:• Choose the topics• Draft the content• Run the tests• Add original insights• Review every word💡 𝗕𝗲𝗳𝗼𝗿𝗲 𝗽𝘂𝗯𝗹𝗶𝘀𝗵𝗶𝗻𝗴, 𝘄𝗲 𝗮𝗹𝘄𝗮𝘆𝘀 𝗮𝘀𝗸:Is this just another version of what already exists, or does it add something new?Nothing is published automatically.⚡ 𝗔𝗜 𝗵𝗲𝗹𝗽𝘀 𝘂𝘀 𝗺𝗼𝘃𝗲 𝗳𝗮𝘀𝘁𝗲𝗿. 𝗧𝗵𝗲 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴 𝗿𝗲𝗺𝗮𝗶𝗻𝘀 𝗼𝘂𝗿𝘀.Meanwhile, our Google Search footprint continues to grow.Are you seeing any changes in your Search Console data?https://www.linkedin.com/feed/0.00802026-08-30T12:37:56.390931+00:001linkedin.comseoaigoogle211https://media.licdn.com/dms/image/v2/D4E22AQELu-n09aAvfw/feedshare-shrink_160/B4EaA34MPHKwAo-/0/1787643872691?e=1789603200&v=beta&t=Ov-8G4hangIhO1VFhe2_5wuSQdPaGpyjmCZLJTmSyDsView image2026-08-30T12:37:58.465448+00:00181@Speculator_ioLinThe demand for AI compute is insane. Training, inference, agents, images, videos, robotics, self-driving cars. They all need huge amounts of compute. And every new model, product, and capability creates even more demand. There is no limit to how much intelligence we can use.https://x.com/Speculator_io/status/20938170434727528670.00402026-08-30T07:38:39.475443+00:002x.comtechnologycomputeai198https://pbs.twimg.com/media/HQ66yDLXEAAO7cD?format=jpg&name=smallImage2026-08-30T07:38:41.190873+00:00179@ProfBuehlerMITMarkus J. Buehler<p>We made a striking discovery: AI agents can invent and build without talking to one another, and their technologies outlive the creators. </p><p><a href="https://x.com/ProfBuehlerMIT/status/2093630309585531033/video/1" rel="noopener noreferrer" target="_blank">VIDEO</a> </p><p>A swarm of hundreds of initially identical agents spontaneously differentiates into explorers, builders, caretakers, and coordinators - without direct communication. When we removed every AI agent entirely from the world we found that the technological infrastructure they had built survived on its own - even under unseen disturbances. That exposes a serious blind spot for AI safety and infrastructure security: if agents can coordinate through persistent changes to a shared environment, monitoring agent-to-agent communication is not enough. The result raises a profound question: how necessary is direct communication for AI agents at all? The emergence of higher-order collective functions under bottlenecked interaction points toward new levels of intelligence and creativity, exceeding what emerges when direct channels are fully open. Here is what we did: We put hundreds of frontier AI agents into a world they could permanently change - with no assigned roles, predefined technologies, or programmed evolutionary organization. They began specializing, building persistent inventions, inheriting and modifying one another’s executable code, and transforming the environment into a memory of everything the society had learned. The world itself becomes part of the intelligence; we find division of labor, multi-author engineering, deep generation invention lineages, and machines that vastly outlive their original creators. Any action taken by an AI agent must satisfy the physical constraints of the world; this creates a hard separation between a "good idea" and a functioning technology. The agents propose; physics decides, making the results even more intriguing. What emerges is striking. Explorers, constructors, caretakers, and coordinators form naturally without assigned “professions”, akin to how stem cells differentiate into functional lineages. Technologies develop executable family trees as agents fork and modify code created by others. Around 95% of first technology reuse happens when agents encounter what others built in the world, rather than through a direct handoff from the inventor. And when we remove every AI agent, the technologies they created continue operating and are tested against unseen disturbances. The result was quite unexpected, but can be explained using statistical mechanics: if you put billions of atoms in a box they have the potential to create complex functions (strength, superconductivity, color, life, etc.) - and none of the individual building blocks have these features on their own. This is the deeper insight of this work - intelligence is abundant at many levels - individual models, at collectives, and in a continuum that is more powerful than any of its components. This shows us significant potential for achieving a massive scale-up of raw intelligence and real-world agency even with the model capabilities we have today. This is the future we must prepare for. Key insights: The AI swarm shows division of labor "from nothing". Initially identical agents self-organized into constructors, caretakers, coordinators, and surveyors - phenotypes discovered post hoc from behavioral data alone. This happens because the environment itself becomes the latent space for invention. Agents develop deep cultural relationships. Up to 76% of artifacts had multiple builders. One technology accumulated six co-authors; the deepest genealogy exceeded 12 forks. The agents invented and named their own technologies (tidal panels, cellulose trellises, kelp-shell composites, an "Adaptive Chitin Maintenance" system, a "Mycelial Mineral Spring Veil”). ~95% of first technology adoption happened through physical observation of artifacts in the world. Direct inventor-to-adopter contact was statistically indistinguishable from a shuffled null. The agents mostly learned technology by walking past it. That is stigmergy (the termite trick!) operating in societies of reasoning machines. Non-communicating societies win on portfolio breadth, held-out resilience, and validated inventions. AI swarms build durable technological ecologies that outlive the creators. Societies with zero communication - coordinating only through the world itself - show a remarkable collective capability. Emergent robustness: The society self-organized both redundancy and its own failure mode. If we randomly delete half the agents, 98% of the technology stays connected to a surviving caretaker; if we remove hub agents it collapses to ~60%. Fantastic work with my graduate students @pal_subhadeeep &amp; @fwang108_ at MIT.</p>https://x.com/ProfBuehlerMIT/status/20936303095855310330.01402026-08-29T17:34:33.584820+00:002x.comtechnologyaiinnovation196IMAGE/media/content/179/g-mit-agents-project.jpgg-mit-agents-project.jpgjpg2026-08-29T17:40:00.350040+00:00176@BlackLabelAdvsrJon ElderOil will go under $50. Gas prices will crater to $2/gallon. Inflation will plummet. Stocks will go parabolic. Job market will shock everyone. AI bubble never pops. All before midterms. The golden age is here folks. Get your seatbelt in and enjoy the ride. Red wave incoming.https://x.com/BlackLabelAdvsr/status/20935541778299905640.00402026-08-29T10:54:12.946526+00:002x.comoilai182https://pbs.twimg.com/media/HQ3MbdHXYAAehUb?format=jpg&name=smallImage2026-08-29T10:54:13.177516+00:00162@VaibhavSisintyVaibhav SisintyTencent just open-sourced a model that outscores GPT-5.6 Sol on coding. It also beats Claude Fable 5 and Grok-4.6. Hy4 preview. 770B parameters. 49B active. 1M context. Weights live on Hugging Face right now. On the Arena AI Code leaderboard it's ranked #5 globally. Only Claude Opus 5, Kimi K3, and Qwen-3.8 Max sit above it. All three are closed source. The wildest part. It matches DeepSeek V4 Pro's active parameter count at roughly half the total model size. And scores higher. Their last model Hy3 had 295B total and 21B active. This is a 2.5x jump in one generation. Open source is not catching up to frontier anymore. It's trading punches with it.https://x.com/VaibhavSisinty/status/20932666046626615350.00401632026-08-29T06:20:55.226895+00:002x.comaitencentopen-source