Practical Techniques for Context Engineering 💡 This is a fantastic blog post from @tuanacelik and @LoganMarkewich on a comprehensive breakdown of the types of context an LLM can interact with, and the core dimensions you have to consider: 1️⃣ Knowledge Base or tool selection -… https://t.co/r6XEtsJpio

In the 87th session of #MultimodalWeekly, we welcome @garridoq_ (Research Scientist at @metaai) to share his awesome paper titled "Intuitive physics understanding emerges from self-supervised pretraining on natural videos" in collaboration with his Meta AI colleagues. https://t.co/lqfZdXHvkM

> a16z funds “Cluely” a startup building AI cheating tools > soham open-sources clone of cluely called “Cheating Daddy” > YC startup steals cheating daddy code and illegally relicenses as Apache 2.0 saying “built in 4 days…”💀 Absolute state. https://t.co/K6WPZomu0e

+2 more

Gemini CLI Update from last week! We merged 85 PRs from 51 unique contributors. Here are key improvements: - Gemini CLI can now use audio and video (santhoshkumarCodes) - Upgrade to Ink 6 and React 19 (SandyTao520) - GEMINI md can import other markdown files with @.… https://t.co/i4CfZKdJOu

lessons from finetuning rerankers with @lancedb https://t.co/54r2a5Cqdj

Whats a minimal viable eval setup? Error Analysis + Notebooks are all you need (for a while) 1 of 3 https://t.co/nv1ukwZ57s

What is "SLURM"? How do you utilize a cluster of GPUs effectively? Why does this matter in today's Deep Learning world? On Wednesday at 11AM EST I'll be talking just about that! https://t.co/AXun5mZd5a

I do not have a favorite eval vendor. This is because I use most of them as a db and build tools on top. What seems to make the most impact wrt to success is the support they provide (which varies according to the situation), so I suggest paying attention to that. https://t.co/ZDNArBEaK1

We're all in on context engineering! A related topic that imo is table stakes for every AI engineer/user: workflow engineering 🛠️ A lot of agent use cases revolve around automating work that otherwise a human would have to perform - customer support, legal research, report… https://t.co/Ry2F1IapZp

I have never been more excited about a talk! Why? @ttorres will show: 1. How domain experts (like PMs) can create high quality evals using simple approaches 2. Solving problems, no gatekeeping. 3. A decisive victory for notebooks. In our course: https://t.co/dR23WB2cAl https://t.co/biFMJgwG6t

now @skylar_b_payne https://t.co/PDgjLQRcuF

What makes a good custom interface for reviewing LLM outputs? (which I recommend most people build!) These are some enhancements we’ve seen work well Screenshots in replies 1/6 https://t.co/W1L8jabX96

In the current AI talent war, everyone is focused on the big numbers (alleged compensation packages). It misses the bigger picture: the cultural shift following the DeepSeek moment. META is the American leader in open science (publications) and open source (Llama). Both OpenAI… https://t.co/5JwjM36Kjk

they really are cooked. adults are not in charge any more https://t.co/gcgzSjuCsJ

RAG POCs are easy, but building production-grade retrieval is legitimately hard. These are things you don’t realize when you’re first starting out building agents - “wow my chat over 10 pdfs works in 10 mins!”. We learned these lessons as we built out LlamaCloud and wanted to… https://t.co/NWPgDrF64x

"Claude 4 Opus, make the most insanely referential thing possible, make it super clever. like really smart. it should be working code" "Make it even more so" https://t.co/OulVDykGyv

Sometimes you get lucky with vibe coding. These days, I rely less on luck and get better results by focusing on context engineering. I built this fully functional deep research agent with Replit Agent and n8n in <10 mins. And it's deployed too! What a time to be alive! https://t.co/kd7K2Kjrb6

Product idea for OpenAI (I know a lot of you follow me): an entirely paper-based LLM. Just 780 volumes and only 30 person years to do the math for the first token using the paper version of GPT-1 Give the weights actual weight. Plus an excellent setup for science fiction stories https://t.co/iDGetnej4H

+1 more

Facebook AI Research (FAIR) is a small, prestigious lab in Meta. We don't train large models like GenAI or MSL, so it's natural that we have limited GPUs. GenAI or MSL's success or failure, past or future, doesn't reflect the work of FAIR. It is important to make this distinction https://t.co/2aN9ZEou7u

How do I evaluate agentic workflows? We recommend a two-phased approach, first do error analysis on end-to-end task success/failure. 1 of 5 https://t.co/ZrfLOuXPWh

buildign a look at your data agent in claude code https://t.co/Igah74qvCB

AI for Scientific Search AI for Science is where I spend most of my time exploring with AI agents. This 120+ pages report does a good job of highlighting why all the big names like OpenAI and Google DeepMind are pursuing AI4Science. Bookmark it! My notes below: https://t.co/z2gRcVbnV4

Threats in LLM-Powered AI Agents Workflows Neat survey of typical threats you encounter when building AI agents. Prompt injections and protocol exploits included. Bookmark this one! https://t.co/WalkxmYRBO

We've partnered with Modular to create Large Scale Inference (LSI), a new OpenAI-compatible inference service. It's up to 85% cheaper than other offerings & can handle trillion-token scale. We originally created it at the request of a major AI lab to do large scale multimodal… https://t.co/Ad6FXWBSXv

Did you know? You can retrieve images and illustrative figures from your LlamaCloud Indexes as well as text! This is great for presentations, reports, and other document types that have rich imagery. Enabling this feature is as simple as toggling the "Multi-modal indexing"… https://t.co/egzmPkOUvv

This is broadly true but the comms here seem very hard to get right “Take a pay cut due to the mission, but as a founder I get to be both mission driven and mega rich” feels like a difficult starting point https://t.co/Er38HO6aym

congrats to Boris and Cat for joining @cursor_ai ! Claude Code + Cursor = ???? https://t.co/oh2q8HkWX6

This absolutely stunning chart comparison by the NYT might prove to be the most important geopolitical visualization of the 21st century. The two major superpowers are each cornering a competing energy platform. China bets everything on clean energy, the US on fossil fuels.… https://t.co/H9mPeFmQy1

sorry apple what in the ever loving fuck is this contact drop down https://t.co/xuCYevJbXK

enough about forward-deployed engineers we need more forward-deployed angels was thinking about this because one of our angels Umesh Khanna (@forwarddeploy) has just been pulling up on Sundays and going deep with us on the product problems we're dealing with, designing growth… https://t.co/HONqYxYwA9

Introducing Document Extraction as an MCP Server ✂️📑 A huge use case for AI agents is being able to extract out items from a diverse set of complex documents in a repeatable manner - whether it’s legal contracts, invoices, financial statements, passports, and more. In this… https://t.co/1glV2lCgZd

Fun! https://t.co/aDbqirMhbw