i meant the older larger non-reasoning models. generally the reasoning models can be pushed to get "there" though, even smaller ones
RESEARCH
-
Inference-Time Compute Improves Reasoning Model Robustness Against Attacks
By
–
Test time continues to be one of the hottest topics this year. "increasing inference-time compute—giving reasoning models more time and resources to ‘think’—can improve robustness to multiple types of attacks…this approach uses reasoning models like o1-preview and o1-mini,
-

Upcoming Chapter Content and Bonus Multi-Head Attention Materials
By
–
Yes, stay tuned! (And there's also the "Comparing Efficient Multi-Head Attention Implementations" bonus content if you are still bored after completing the chapter :))
-
Building LLM Agents with Computer Use Capability
By
–
Our first short course with @AnthropicAI! Building Towards Computer Use with Anthropic. This teaches you to build an LLM-based agent that uses a computer interface by generating mouse clicks and keystrokes. Computer Use is an important, emerging capability for LLMs that will let… pic.twitter.com/ghG3uGcRuP
— Andrew Ng (@AndrewYNg) 22 janvier 2025Our first short course with @AnthropicAI
! Building Towards Computer Use with Anthropic. This teaches you to build an LLM-based agent that uses a computer interface by generating mouse clicks and keystrokes. Computer Use is an important, emerging capability for LLMs that will let -
Making Data Freely Available to LLMs for Training
By
–
Explicitly and eagerly available to LLMs, for free, for training and RAG, and start the movement.
-

Workflows versus Agents: Understanding AI Collaboration and Decision-Making
By
–
Think of workflows as task-doers—great for answering questions about a database, for example. With Agents, we aim to create collaborators who can think (plan and iterate with themselves) and make decisions, as in this great illustration by Anthropic:
-

Real AI Agents Beyond Predefined Workflows and Hard-Coded Solutions
By
–
These systems aren’t independent or intelligent; they are predefined processes and hard-coded solutions called “workflows”, as seen below. Real agents, like Anthropic describes in their awesome recent blog, dynamically control their own tasks, reason, reflect, and adapt without
-
AGI Advancement Outpacing Human Readiness and Business Adaptation
By
–
The delta between AI advancement and human readiness is growing too fast, and no one seems incentivized enough to close it. We’re hurtling toward AGI, but businesses are barely keeping up with GenAI. Compare the following… Anthropic CEO predicts AGI (or what he calls “a
-
AI Solving Major Hard Science Challenges Beyond Capex
By
–
It literally is solving the biggest challenges in the hard sciences im not even talking about capex. But you do you.
-
Tech Investment in AI: AlphaFold Proves Real Value Beyond Hype
By
–
Cool, I guess those ‘tech bros’ just pumped in 500bn (without considering value prop) for … jokes? AlphaFold ALONE is enough proof of value. C’mon. Don’t be the IBM guy who said … “there is a world market for maybe five computers.”