Jeffrey Zhao

@GoogleDeepMind

[Andrej Karpathy @karpathy]

Judging by my tl there is a growing gap in understanding of AI capability.

The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.

But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.

So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.

TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

[Ali @AliKh1886]

When Trump discovered ChatGPT:

[Sundar Pichai @sundarpichai]

Introducing Gemini 3 ✨

It’s the best model in the world for multimodal understanding, and our most powerful agentic + vibe coding model yet. Gemini 3 can bring any idea to life, quickly grasping context and intent so you can get what you need with less prompting. 

Find Gemini 3 Pro rolling out today in the @Geminiapp and AI Mode in Search. For developers, build with it now in @GoogleAIStudio and Vertex AI. 

Excited for you to try it!

[Saeejith Nair @sighjith]

I built https://t.co/VJ7oHdvf7s, a new interface for arXiv

As we enter an era of accelerated scientific discovery, we need better tools that augment human cognition to help us keep up.

Try it: visit papiers ai or swap arxiv -> papiers on any paper URL

[Donald J. Trump @realDonaldTrump]

What separates the winners from the losers is how a person reacts to each new twist of fate.

[Google DeepMind @GoogleDeepMind]

Deep Think in 2.5 Pro has landed. 🤯

It’s a new enhanced reasoning mode using our research in parallel thinking techniques - meaning it explores multiple hypotheses before responding.

This enables it to handle incredibly complex math and coding problems more effectively.

[Google DeepMind @GoogleDeepMind]

We’ve developed Gemini Diffusion: our state-of-the-art text diffusion model.

Instead of predicting text directly, it learns to generate outputs by refining noise, step-by-step. This helps it excel at coding and math, where it can iterate over solutions quickly. #GoogleIO

[Thomas Woodside 🫜 @Thomas_Woodside]

The Pope actually chose the name Leo because of AI

[Noam Shazeer @NoamShazeer]

Introducing Gemini 2.5 Pro Experimental.

The 2.5 series marks a significant evolution: Gemini models are now fundamentally thinking models.

This means the model reasons before responding, to maximize accuracy -- and it’s our best Gemini model yet.

Blog - https://t.co/Gx35wlkomq

[Jeffrey Zhao @jezhao]

@SchmidhuberAI @DjokerNole 😳

[Jeffrey Zhao @jezhao]

@yongyuanxi I never ironically liked Panino's. Mikey's Eatery was the better bad restaurant

[Jeffrey Zhao @jezhao]

@sama what

[Jeffrey Zhao @jezhao]

@YiTayML Doing this all in 6 months is incredible… congrats!!

[Jeffrey Zhao @jezhao]

@_arohan_ gopher snake?

[AK @_akhaliq]

Tree of Thoughts: Deliberate Problem Solving with Large Language Models

experiments show that ToT significantly enhances language models' problem-solving abilities on three novel tasks requiring non-trivial planning or search: Game of 24, Creative Writing, and Mini Crosswords. For instance, in Game of 24, while GPT-4 with chain-of-thought prompting only solved 4% of tasks, method achieved a success rate of 74%

paper page: https://t.co/CoBjFTBIg4

[Shunyu Yao @ShunyuYao12]

Still use ⛓️Chain-of-Thought (CoT) for all your prompting? May be underutilizing LLM capabilities🤠

Introducing 🌲Tree-of-Thought (ToT), a framework to unleash complex & general problem solving with LLMs, through a deliberate ‘System 2’ tree search.

https://t.co/V6hjbUNjbt

[Jeffrey Zhao @jezhao]

@spiantado I guess the intuition is that, since there are finite colors, there is a color that will have infinitely many n with diverging sum. Infinitesimally small 1/n -> you can always find some m in your color such that your current sum + 1/m <= 1 but closer to it.

[Google AI @GoogleAI]

Learn how ReAct, a general paradigm for synergizing reasoning and acting in language models, presents more interpretable, diagnosable, and controllable task-solving trajectories while outperforming reasoning and acting only paradigms. Read the blog → https://t.co/BZg2VzqBZk

[Shunyu Yao @ShunyuYao12]

Large Language Models (LLM) are 🔥in 2 ways:
1.🧠Reason via internal thoughts (explain jokes, math reasoning..)
2.💪Act in external worlds (SayCan, ADEPT ACT-1, WebGPT..)
But so far 🧠and💪 remain distinct methods/tasks...

Why not 🧠+💪?
In our new work ReAct, we show 1+1>>2!

[Google AI @GoogleAI]

Task-oriented dialogue models power many modern conversational agents, but can be limited in their ability to generalize to new tasks. Check out two new approaches for dialogue modeling that use additional context during training for improved performance. https://t.co/IsY5nRIbCX