<?xml version='1.0' encoding='utf-8'?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" xmlns:media="http://search.yahoo.com/mrss/" version="2.0"><channel><title>Gnosly Feed - self-driving</title><link>https://gnosly.com/feed/self-driving/</link><description>Business Intelligence Console and Prediction Algorithms</description><atom:link href="https://gnosly.com/feed/self-driving/" rel="self" type="application/rss+xml" /><item><guid isPermaLink="false">gnosly:4250</guid><title>Waymo puts about $200,000 into every robotaxi. Tesla's Cybercab is going after the same rides with 8 cameras and a $30,000 price tag. If Tesla's camera-only bet works, robotaxi fleets get about 85% cheaper to build.</title><description><![CDATA[<p>Since Tesla opened Cybercab rides in Austin, the comparison is wild:</p><p>1️⃣ Sensors: Waymo runs 29 cameras, 5 LiDARs, and 6 radars, while Tesla runs 8 cameras and AI only</p><p>2️⃣ Cost: about $200K per Waymo (Morgan Stanley estimate) against about $30K for a Cybercab</p><p>3️⃣ Inside: a Waymo is a regular Jaguar, while the Cybercab drops the steering wheel and pedals for more space and a giant screen. </p>]]></description><link>https://lnkd.in/p/dMZYxG7v</link><pubDate>Sun, 15 Nov 2026 10:21:50 +0000</pubDate><category>robotaxi</category><category>tesla</category><category>ai</category><category>self-driving</category><enclosure url="https://media.licdn.com/dms/image/v2/D4E22AQEdxdZ8mQ0teg/feedshare-shrink_800/B4EaHea7rKKgAg-/0/1794732964840?e=1796256000&amp;v=beta&amp;t=_Wu6HFzi9_L76SreCcMq25sNqGfDdIi12tpXnHdpouE" length="0" type="image/jpeg" /><media:content url="https://media.licdn.com/dms/image/v2/D4E22AQEdxdZ8mQ0teg/feedshare-shrink_800/B4EaHea7rKKgAg-/0/1794732964840?e=1796256000&amp;v=beta&amp;t=_Wu6HFzi9_L76SreCcMq25sNqGfDdIi12tpXnHdpouE" medium="image" type="image/jpeg" /></item><item><guid isPermaLink="false">gnosly:4240</guid><title>NVIDIA just released Alpamayo 2 Edge, a compact Vision-Language-Action (VLA) model for on-device autonomous driving, shrinking its autonomous driving AI from 34B to 2B parameters so it can run on the edge. That's a 17× reduction in model size, bringing AI reasoning and driving decisions directly onto edge hardware.</title><description><![CDATA[<p><strong>Here are the details</strong> on this tiny but packed model: </p><p>- Reasoning on the Edge: Interprets road conditions, intersections, and driving intent before generating trajectories. </p><p>- Diffusion-Based Planning: Generates 6 candidate driving trajectories with 64 waypoints over a 6.4-second horizon. </p><p>- Multi-Camera Perception: Processes 6 camera streams with vehicle motion data. </p><p>- Lightweight Inference: Continuous evaluation on a free Google Colab T4 GPU. </p><p><strong>The main takeaway?</strong> </p><p>The shift is from massive autonomous driving models to compact models that can reason and plan directly on edge devices. 17× smaller. Intelligence on the edge. A big step for Physical AI.</p>]]></description><link>https://lnkd.in/p/dh-Sie6P</link><pubDate>Sun, 15 Nov 2026 06:12:15 +0000</pubDate><category>nvidia</category><category>physical-ai</category><category>ai</category><category>self-driving</category><enclosure url="https://media.licdn.com/dms/image/v2/D4D05AQFyd3L0KrPArA/videocover-low/B4DaHYz95uIMBM-/0/1794638865248?e=1795330800&amp;v=beta&amp;t=xru1KjNP403lEsTpkeul-Pv5it-TfMCFXDlA_xRtpGw" length="0" type="image/jpeg" /><media:content url="https://media.licdn.com/dms/image/v2/D4D05AQFyd3L0KrPArA/videocover-low/B4DaHYz95uIMBM-/0/1794638865248?e=1795330800&amp;v=beta&amp;t=xru1KjNP403lEsTpkeul-Pv5it-TfMCFXDlA_xRtpGw" medium="image" type="image/jpeg" /></item></channel></rss>