Anthropic @AnthropicAI ❤️ zkSTARKs:
In collaboration with @StarkWareLtd co-founder Prof. Alessandro Chiesa and other brilliant researchers, Anthropic just open sourced a zk STARK system for proving the integrity of LLM inferencing, bringing us closer to verifiable AI.
As we said long ago (since 2018), zkSTARKs are the end-game: highest scale, best security (post-quantum). Now, also AI agile and supporting AI integrity.
https://t.co/kgAMnOVhtw
Esto se va a poner peor, les cuento:
Lo de $INTC es el resultado de que nadie quiere comprar un computador mucho más caro por una memoria RAM que cuesta hasta 5 veces más que hace un año, un precio que no lo vale. El volumen de ventas de PC de Intel cayó 14% y los ingresos del segmento bajaron 5%.
Reddit no ve un futuro fuerte porque los agentes le están quitando el tráfico que viene de Google. Más de 1 millón de referencias de búsqueda perdidas por semana en octubre y advierten que sus usuarios diarios podrían caer a doble dígito.
Google se pegó la hostia del año la razón? AI Mode está generando menos ingresos publicitarios por búsqueda. El crecimiento de Search cayó del 17% al 12% porque la IA monetiza menos que el buscador tradicional. Los agentes están cambiando el negocio de la publicidad y Google ya está sintiendo el golpe.
Microsoft está ofreciendo salidas voluntarias a unos 60.000 empleados mientras acelera la automatización de tareas con agentes de IA.
El problema, mis queridos compañeros, es que estamos viendo un cambio sin precedentes a nivel económico del que nadie quiere hablar.
Esto a mi parecer solo va a empeorar y ojo, que la cereza del pastel son los hackeos con IA y vulnerabilidad.
Mientras el mercado se adapta muchas cosas van a complicarse, muchos negocios van a sufrir, sin contar con la optimización de la inteligencia artificial que tumba expectativas de crecimiento en hardware también...
Vendrán oportunidades, pero hay que ser pacientes.
We’re disclosing HEIF Heist, a months-long investigation into libheif that allowed us to hack OpenAI, Slack, Meta, GitHub Ent, Rails, Next.js, ImageMagick, and many more.
It was literally xkcd #234, one obscure image library beneath a huge number of apps. 🧵
We gave a fly $100 to trade on Polymarket.
It already called today’s pump in the "25 bps increase" on Fed Decision polymarket.
Every second, its next move is decided by 166,700 neurons.
Today we published WeWorm, our zero-click worm that spreads across iOS and Android.
All it takes is one phone call. You don't have to answer. Seconds later, your WeChat account is compromised, calling your friends and spreading the attack.
We reported the bug to Tencent, and it's now mitigated for all users.
We hope this sets an example. The US and China disagree on plenty, but keeping billions of people safe online shouldn't be one of them.
AI gives us a chance to find and fix these bugs faster than ever. We should work together to make the world safer for everyone.
Our story and demos: https://t.co/YsoYFduv60
No. I still don't think that pure scaling of LLMs can lead to AGI.
BUT...
... most recent advancements aren't actually about scaling LLMs. The really about adding deterministic, symbolic stuff to LLMs.
Even that is not sufficient to get to AGI.
But it is necessary. (see my article Next Decade in AI).
1/ A major breakthrough on the Riemann Hypothesis recently shocked the world: >67% of zeta zeros lie on the critical line and are simple.
Now Youness Lamzouri has found a strikingly elegant new proof.
And AxiomProver has already formally verified it, in his paper’s appendix.🧵
This is the intelligence plateau sign everyone's been looking for.
5.6-sol and Fable have reached the intelligence of the smartest humans. There is nothing left to scale toward because there is no higher intelligence to learn from.
IQ scaling is done. Progress from here is benchmark hacking and personality wars. Models will stop looking like their CEOs. Opus won't always be a bitch like Dario. Some Grok models will be less of a troll than Elon.
Because of the current progress in AI, many people are starting to think that the Riemann hypothesis could be solved ''soon''!
In the meantime, I suggest you to read this fantastic presentation (40 pages) by the legend Alain Connes, titled 'The Riemann Hypothesis: Past, Present and a Letter Through Time', a (technical) history of the Riemann hypothesis (and more).
🔗👇