In era of pretraining, what mattered was internet text. You'd primarily want a large, diverse, high quality collection of internet documents to learn from.
In era of supervised finetuning, it was conversations. Contract workers are hired to create answers for questions, a bit…
In era of pretraining, what mattered was internet text. You'd primarily want a large, diverse, high quality collection of internet documents to learn from.
In era of supervised finetuning, it was conversations. Contract workers are hired to create answers for questions, a bit…
+1 for "context engineering" over "prompt engineering".
People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window…
+1 for "context engineering" over "prompt engineering".
People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window…
From my friend @vinayramasesh:
"I would benefit most from an explanation style in which you frequently pause to confirm, via asking me test questions, that I've understood your explanations so far. Particularly helpful are test questions related to simple, explicit examples.…
We're missing (at least one) major paradigm for LLM learning. Not sure what to call it, possibly it has a name - system prompt learning?
Pretraining is for knowledge.
Finetuning (SL/RL) is for habitual behavior.
Both of these involve a change in parameters but a lot of human…
We’re bringing 50+ languages to @NotebookLM’s Audio Overviews, including French, Hindi, Japanese + Portuguese, with more on the way. 🌍 Get podcast conversations based on the sources you give it, even if they are in different languages. Our Audio Overview hosts demonstrate:
PSA It’s a new era of ergonomics.
The primary audience of your thing (product, service, library, …) is now an LLM, not a human.
LLMs don’t like to navigate, they like to scrape.
LLMs don’t like to see, they like to read.
LLMs don’t like to click, they like to curl.
Etc etc.
PSA It’s a new era of ergonomics.
The primary audience of your thing (product, service, library, …) is now an LLM, not a human.
LLMs don’t like to navigate, they like to scrape.
LLMs don’t like to see, they like to read.
LLMs don’t like to click, they like to curl.
Etc etc.
Gemini Advanced is now free for college students in the U.S.
The best part: This offer isn't just for finals this semester. Sign up before June 30, 2025, and you’ll get free access all the way through spring finals 2026.
Terms apply.
The New York Times is pure propaganda.
Also, I look forward to the prosecutions of those at the Pentagon who are leaking maliciously false information to NYT.
They will be found.
When working with LLMs I am used to starting "New Conversation" for each request.
But there is also the polar opposite approach of keeping one giant conversation going forever. The standard approach can still choose to use a Memory tool to write things down in between…
Seeding my Bear ʕ•ᴥ•ʔ blog with more random posts, e.g. here's something I had on backlog for a while:
# The append-and-review note
An approach to note taking that I stumbled on and has worked for me quite well for many years. I find that it strikes a good balance of being…
It's 2025 and most content is still written for humans instead of LLMs. 99.9% of attention is about to be LLM attention, not human attention.
E.g. 99% of libraries still have docs that basically render to some pretty .html static pages assuming a human will click through them.…
New 2h11m YouTube video: How I Use LLMs
This video continues my general audience series. The last one focused on how LLMs are trained, so I wanted to follow up with a more practical guide of the entire LLM ecosystem, including lots of examples of use in my own life.
Chapters…
This is interesting as a first large diffusion-based LLM.
Most of the LLMs you've been seeing are ~clones as far as the core modeling approach goes. They're all trained "autoregressively", i.e. predicting tokens from left to right. Diffusion is different - it doesn't go left to…
This is interesting as a first large diffusion-based LLM.
Most of the LLMs you've been seeing are ~clones as far as the core modeling approach goes. They're all trained "autoregressively", i.e. predicting tokens from left to right. Diffusion is different - it doesn't go left to…
This is moralistic garbage, which is unfortunately the rhetorical currency of the globalists because they have nothing else to say.
For three years, President Trump and I have made two simple arguments: first, the war wouldn't have started if President Trump was in office;…
This is moralistic garbage, which is unfortunately the rhetorical currency of the globalists because they have nothing else to say.
For three years, President Trump and I have made two simple arguments: first, the war wouldn't have started if President Trump was in office;…
167 Followers 138 FollowingBuilding the future of behavioral AI in games @ArtificialAgncy | powersports enthusiast | fmr. @coalitiongears @improbableio @bioware @ea
145 Followers 3K FollowingAm living with my siblings our both parents has passed away few years ago we are looking for help at all over the world. Out mum live a 2 year old young baby🍲
342 Followers 493 FollowingExploring the latest breakthroughs in Artificial Intelligence | Uncovering the future of AI technology and its impact on society 🤖
334K Followers 182K Following⚡️ Bringing daily magic to your timeline with fun facts, memes and everything Harry Potter related! 🪄 By fans for the fans. Independent of Warner Bros ✨
20K Followers 5 FollowingImpossible? Let’s see. From algorithms to neuroscience to AI, Google Research strives to progress science, advance society & improve billions of people’s lives.
8K Followers 3K FollowingCo-founder & President @NewLimit | Prev: PI @Calico, PhD @UCSF | Interested in aging, machine learning, genomics, the technology production function
9K Followers 13 FollowingHi👋 We're Flow, an AI filmmaking tool from Google, powered by Veo. Follow along for product news and creative tips to bring your cinematic stories to life.
197K Followers 38 FollowingThe Gemini app turns research into reality, bringing frontier AI experiences like Veo 3, Deep Think, and more to hundreds of millions of people.
204K Followers 624 FollowingFPS is my DNA. Game Director @RomeroGamesLtd | New FPS in the works | DOOM, DOOM II, Quake, Wolfenstein 3D, Heretic, Hexen, SIGIL, SIGIL II.
26K Followers 132 FollowingCreated Apogee (1990) and Apogee's "3D Realms" label (1994). Also co-created Duke Nukem, Max Payne, Prey and others. Follow to see game history revealed!
101K Followers 43 FollowingBuilding the Android of self-driving cars.
comma 3X is available now for $999, plugs into the car you already drive, and drives half your miles.
45K Followers 44 FollowingActive on https://t.co/WG71Nrs60M; also trying out https://t.co/fGOzbSxVHi. No longer read replies or notifications here now that tweetdeck is gated.
17K Followers 181 FollowingChief Quant, Dir Analytics (+many hats) @_MathAcademy_.
I do our algo-heavy infra including the AI expert system & knowledge graph.
✍️ about serious upskilling
28K Followers 42 FollowingOklo’s mission is to provide clean, reliable, affordable energy through the deployment of next-generation fast fission technology and recycling nuclear waste.
3K Followers 1K FollowingGenerative AI Product leader at @GoogleDeepMind working on Gemini API and AI Studio @Google; Previously @Theteamatx, @Salesforce @Docusign
Opinions my own
2K Followers 342 FollowingPartner @kleinerperkins. Love building teams. Seen every mob movie there is. Told I have a face for radio: https://t.co/cnpC9jTTVK
180K Followers 4K FollowingWriting at https://t.co/m6EtO60SiY and host of the Core Memory podcast. 2X NYT best-seller. Filmmaker @HBO (Wild, Wild Space) + @Netflix (Don't Die).
9K Followers 102 FollowingMember of Technical Staff at Anthropic AlphaGo, AlphaZero, MuZero, AlphaCode, AlphaTensor, AlphaProof Gemini RL Prev Principal Research Engineer at DeepMind
6K Followers 5K FollowingAI is good & bad, actually.
Tweeting about AI/ML methods, software dev, research, tech and society, social impact.
20yrs in tech, 10 in ML/AI, PhD in comp sci
3K Followers 1K FollowingProfessor of Physics & Senior Data Fellow @BelmontUniv, teaching Audio Engineers. Head of Research @Hyperstate_AI. Mostly: ML for music producers.
1.1M Followers 1 FollowingPAC Founded by @ElonMusk to support candidates who champion Secure Borders, Sensible Spending, Safe Cities, Fair Justice System, Free Speech and Self-Protection
18K Followers 334 FollowingCofounder @Newlimit to extend our lifespans. Investor in early stage ambitious founders, 10 IPOs and counting. ex @GVteam. @Stanford bioengineering PhD.