lunes, 2 de noviembre de 2026

Andrej Karpathy @karpathy

Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5: https://hackertube.org/ - Excellent recs without any logins (state is kept in local browser session) - Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads - Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version - HackerTube Longs ✊ A bit more on the process: I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing. Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.

Andrej Karpathy muestra cómo la IA de próxima generación puede eludir feeds algorítmicos para crear plataformas de contenido a medida.

unusual_whales @unusual_whales

BREAKING: Anthropic's Claude AI blocks video game cloning and reverse-engineering, per Toms Hardware

Anthropic restringe a sus modelos de replicar juegos con derechos de autor, destacando la tensión entre la codificación de IA y la PI.

Chetaslua @chetaslua

Fable 5.5 its over for video and image models 🤯 THIS IS PURE CODE one shotted this video holy shit , how this is more beautiful than ai video models Two ways to check this yourself: Ask: “Where did @karpathy go in 2026?” with web search and memory turned off. Only the new Fable knows the answer. Then give it the prompt. Run the same prompt I used for the test (it’s in the screenshot) and see for yourself. Don’t rely on Grok for basic reasoning like this. Grok doesn’t have up-to-date knowledge of niche stuff. Either join the community or run the test yourself. Also, a knowledge cutoff doesn’t mean a model knows everything before that date. Sometimes a model doesn’t know about an event even if it happened before its cutoff. This test shows that too. Run the same prompt on regular Opus 5.5 and the quality drops. On Fable 5.1 it’s worse.

Un nuevo modelo de IA llamado Fable 5.5 genera videos complejos directamente desde código, una alternativa a los modelos de difusión.

Sawyer Merritt @SawyerMerritt

Tesla has just officially announced a new Robotaxi city, their first new city expansion in ~3 months! Fort Worth, Texas Tesla robotaxi rides begin on November 2nd at 6 AM CT.

Tesla expande su red de viajes autónomos a Texas, marcando su primera nueva ciudad operativa para el servicio en tres meses.

Sawyer Merritt @SawyerMerritt

With today’s launch of @Tesla’s Robotaxi service in Fort Worth, Texas, its unsupervised Robotaxi geofences now cover a combined ~1% of the entire U.S. population.

El lanzamiento en Fort Worth impulsa la red de viajes sin conductor de Tesla más allá de un gran hito demográfico en Estados Unidos.

Wall St Engine @wallstengine

$MRVL TO ACQUIRE $LITE IN $115B ALL-STOCK DEAL Marvell has agreed to acquire Lumentum in an all-stock transaction valued at approximately $115B in enterprise value, combining two major suppliers of AI infrastructure and optical networking technology. Under the agreement: • Lumentum shareholders will receive 4.5 Marvell shares for each LITE share, implying roughly $1,132 per share. • The offer represents a 23.9% premium to unaffected closing prices on October 28. • Marvell shareholders will own approximately 67% of the combined company, with Lumentum shareholders holding 33%. The deal brings together Marvell's custom AI chips, networking switches and interconnect technology with Lumentum's lasers, optical transceivers and optical circuit switching capabilities. The bigger opportunity is the growing shift from copper to optical connectivity as AI data centers require faster data transfer, greater bandwidth and improved power efficiency. Marvell says the combination could expand its addressable market to $550B, allowing it to offer more integrated connectivity solutions across AI data centers and within advanced chip packages. The companies also expect: • More than $1B in annual run-rate cost synergies. • Improved long-term revenue growth and profit margins. • Accretion to Marvell's adjusted EPS in the first year after closing. Matt Murphy will remain CEO of the combined company, with the transaction expected to close in H1 2028, subject to shareholder and regulatory approvals. Separately, Marvell reaffirmed the midpoint of its fiscal Q3 revenue guidance while narrowing the expected range to ±2% from ±5%.

Marvell consolida la cadena de suministro de hardware de IA con la adquisición masiva del gigante de redes ópticas Lumentum.

prinz @deredleritt3r

Sam Altman says that the internal model that solved Navier-Stokes will be released "soon": "It was 10,000 agents of the model that is going to be released soon, the next GPT, working for about four days, nonstop, collaborating with each other perfectly." https://youtu.be/2YWfVGpVtvc?t=332

OpenAI confirma que su próximo modelo incluirá el razonamiento multiagente usado recientemente para resolver dinámica de fluidos compleja.