I would love to see a highlight-to-explain feature in LLMs to talk about part of a message directly.
It happens often that it mentions an unfamiliar term/concept. I should be able to do highlight -> explain and have it open a small side-window with an explanation for it and how…
I've found that GPT5 is able to find the most subtle bugs if I ask it to think hard and only stop when it has its bingo moment. Takes a few minutes but its getting them right every single time.
Finally. An AI lab has figured out that you should put a viewfinder in the chat interface. Hope everyone copies them fast cause IDK why this did not exist already.
My personal Twitter guidelines (absolutely not a 🧵):
• Encourage other people - if you like the work someone is doing, tell them!
• Be positive - a feed full of negativity is zero fun to follow.
• Share what you're working on - people are drawn to other people in motion.
You know when you're on a roll and doing the thing right every day like you're supposed to but then you fall off the bandwagon for some reason or the other?
How do you get back to it then after failing to do it properly for a while?
I seem to find my way back to doing the…
I'm extremely bullish on using AI to learn new skills at a faster speed than you could ever before so I'm building an app which uses AI to teach you how to sing.
Thoughts on the landing page?
If some mech interp guys manage to prove that LLMs have some inherent limitation that won't let it scale to AGI then are they just going to be killed by the BIG AI?
31K Followers 4K Following📙 Cracking the PM Interview 📒 Cracking the PM Career 👩🏼💻 Previously Head of Product Management @ Asana, PM @ Google & Microsoft.
83K Followers 8K FollowingCompiling in real-time, the race towards AGI.
🗞️ Don't miss my daily top 1% AI analysis newsletter directly to your inbox 👉 https://t.co/6LBxO8215l
254 Followers 441 FollowingFounder @Greipio — Backing top 100 US financial institutions and global ecoms with fraud tools delivering 10x-20x ROI on average
436 Followers 560 FollowingBIOS keyboard interrupt
Meta Research Scientist (5+ yrs) | ML / CV / CG
Writing CUDA kernels for differentiable rendering
PhD CS • MS Aerospace Eng
Ex-gamedev
42K Followers 865 FollowingFR/US/GB AI/ML Person, Director of Research at @GoogleDeepMind, Honorary Professor at @UCL_DARK, @ELLISforEurope Fellow. All posts are personal.
4K Followers 220 FollowingAI Researcher: From theory to practice (and back)
Postdoc @MetaAI with @ylecun
PhD @RiceUniversity with @rbaraniuk
Masters @ENS_Ulm @Paris_Sorbonne
42K Followers 109 Following• Center for AI Safety Director
• xAI and Scale AI advisor
• GELU/MMLU/MATH/HLE
• PhD in AI
• Analyzing AI models, companies, policies, and geopolitics
46K Followers 1K FollowingWriter https://t.co/TquuQXlLOJ. O'Reilly Author https://t.co/Fl3uPAZHLg. LLM Builder @Cohere. Visualizing AI one concept at a time.
8K Followers 636 Followingbuilding @flowith_ai to shape the interface for AGI. tech optimist, AI engineer, designer. previously founded X ACADEMY and Realm ($2M ARR).
5K Followers 175 FollowingIncoming Assistant Professor of Machine Learning @CarnegieMellon, Research Scientist at Google DeepMind, Berkeley PhD @Berkeley_AI
10K Followers 1K Followingdesign engineer @netflix, writing and diving into my curiosities @rabbitholeathon, investing in creative tools! uw syde grad 🇨🇦
5K Followers 708 Followingassociate prof @ucberkeley, co-director @ucbepic, cofounder @ponderdata (acq. @snowflakeDB) | on a mission to make data science effortless at scale | he/him
30K Followers 123 FollowingMechanistic Interpretability lead DeepMind. Formerly @AnthropicAI, independent. In this to reduce AI X-risk. Neural networks can be understood, let's go do it!