Shane Legg

Project organization is here: Introducing notebooks in Gemini. You can now keep multiple projects organized and even add past chats and relevant files as sources, so you have a dedicated space to focus on the task at hand. Select “New notebook” in the side panel to get started.

@ishanhaq See the end of this post from 2009 vetta.org/2009/12/tick-t…

Woah. @ShaneLegg's prediction 14 years ago... 🎯

It’s happened. Mac Studio is here. Gemma 4 31b @GoogleDeepMind installed, chatting with my main @openclaw for $0 in token expenses now... I've burned $5-6k on tokens on my crazy ideas over past few months, so this mac studio should pencil out for me within 3 months or so 🤓

My team (Frontier Strategy & Governance) at Google DeepMind is hiring. We are bringing on researchers to lead our strategic analysis on frontier AI. Our mission is to provide rigorous foresight and actionable insights to help the world prepare for advanced AI. We focus on concrete, longer-horizon challenges that fall outside the ordinary course of business: forecasting technical breakthroughs, anticipating geopolitical dynamics, contributing to the development of international safety standards, and planning scientific moonshots. We also publish external research and convene experts. Recent public work covers lessons from historical technological revolutions, forecasting frameworks for frontier AI, Cooperative AI, and the limits of model-level governance. This builds on our team's track record of co-creating the Frontier Model Forum, the Frontier Safety Framework, and the AGI Safety Council. We are seeking candidates with deep expertise in domains such as AI governance, international security, political economy, compute governance, forecasting, and institutional design. You will need the analytical skill to translate complex technical and political trends into internal briefings, external publications, and company initiatives. Experience briefing executive leadership is highly valued. London is preferred, with hubs in NYC and the Bay Area. We are targeting mid-career and senior researchers, though exceptional early-career applicants are considered. Apply here: job-boards.greenhouse.io/deepmind/jobs/…

Wonderful trip to Japan with @ShaneLegg & @shanegJP. It's really refreshing to experience such a different and open cultural/philosophical attitude towards AGI. Less reactive alarmism, more openness to positive futures, a more adaptive regulatory environment, and a good dose of realism given the demographic crisis. Will definitely be back! 🗻

@azi_pat @kaggle See end of this article from 2009: share.google/QRwzzzGZfa5jGK…

@ShaneLegg @kaggle Define “in the coming years.” What’s your timeline?

Check out this great work on measuring progress towards AGI and the associated global @Kaggle hackathon. I continue to believe that Minimal AGI will be achieved in the coming years: an AI that can do all the cognitive things that people can typically do.

this is so fucking wholesome guy used AI to save his cancer-ridden dog by sequencing its DNA and creating a CUSTOM cure. the tech behind this is fucking awesome (well done @demishassabis and the google team): - used CHATGPT to sequence dogs DNA discovers mutations - ran the mutations through Google’s Alphafold (AI protein sequencer) which CREATED A CUSTOM VACCINE TO TREAT THEM. - treated dog and reduced tumour by 50% in WEEKS. dog is alive and well. - this is the 1st time AI has been used to create a custom vaccine for a dog (and it worked) - dude is now working on similar vaccines for humans using AI! 2026 is definitely the year we see AI change personalised medicine in a HUGE way so sick

Almost every successful person begins with two beliefs. “The future can be better than the present, and I have the power to make it so.” — @nytdavidbrooks

Our GenAI traffic share update is back. This chart tracks the monthly traffic share of leading AI tools worldwide. Main takeaways: → As of February, Grok and Claude surpassed DeepSeek, taking 3rd and 4th place respectively. → Claude crossed the 3% mark for the first time in February. → Gemini is approaching a quarter of the total share. 🗓️ 12 months ago: ChatGPT: 75.7% DeepSeek: 8.5% Gemini: 5.7% Grok: 3.4% Perplexity: 2.1% Claude: 1.7% Copilot: 1.3% 🗓️ 6 months ago: ChatGPT: 74.0% Gemini: 13.3% DeepSeek: 4.2% Grok: 2.2% Perplexity: 2.1% Claude: 2.0% Copilot: 1.2% 🗓️ 3 months ago: ChatGPT: 65.8% Gemini: 20.7% DeepSeek: 3.9% Grok: 3.2% Perplexity: 2.1% Claude: 2.1% Copilot: 1.2% 🗓️ 1 month ago: ChatGPT: 61.7% Gemini: 24.4% Grok: 3.4% Claude: 3.3% DeepSeek: 3.2% Perplexity: 1.8% Copilot: 1.1% >>

Google has shipped a CLI for Google Workspace (Drive, Gmail, Calendar, Sheets, Docs, …) Huge! Written in Rust, distributed through npm & skills.sh $ npm i -g @⁠googleworkspace/cli $ npx skills add github:googleworkspace/cli 2026 is the year of Skills & CLIs github.com/googleworkspac…

Layers upon layers upon layers......

ai progress has stalled no major ai news stories since breakfast

AI training compute efficiency has improved extremely fast: each year, you need several times less training compute to reach the same capability. But AI architectures/algorithms haven’t changed *that* much in recent years. So where do these efficiency improvements come from? 🧵

@HugeneOrmus @bryan_johnson Sometimes I travel and don't look at it for a couple of weeks. Not a problem.

@ShaneLegg @bryan_johnson You won't know if you experience this until you stop for a month, that is the entire point of Bryan's post

Social media has started feeling repulsive to me after these fasts. The time away broke the dopaminergic spell. I can now feel what I was numb to. It feels like brain rot, a blood curdling sound and assault. This is complicated. I really enjoy posting and the interaction.

@wlgriffiths @bryan_johnson 40 minutes.

@ShaneLegg @bryan_johnson How much time per day are you on X or other socials?

New paper! It’s now well known that Frontier AI models are still quite jagged, e.g. outperforming expert mathematicians while failing at seemingly easy visual puzzles. We argue that it’s worth better characterizing and measuring the jaggedness of new models.

Google is now dominating ARC-AGI-2 with Gemini 3 Flash, Gemini 3.1 Pro and Gemini 3 Deep Think (Feb)

@AnnaLeptikon youtu.be/5FON7aMlKQM?t=…

That's gonna be me once I am healthy again. Just running around some green landscape in a dress grinning like a maniac.