AI is speeding up. 7x up in just this instance. Today's lesson from NVIDIA's GTC? AI factories are generating tokens faster and cheaper than last year and that's making new things possible. I'm feeling it, as I use AI to make a new X news site that we'll turn on this weekend.
GENERATIVE AI
-
Tokens-per-watt: New AI Productivity Measurement Standard
By
–
‘Tokens-per-watt’ is the new measure of productivity. It’s not a ‘tool’ a human uses, it’s a ‘capability’ run by your AI agent.
-
AI Shifts Bottleneck from Code Generation to Code Review
By
–
The bottleneck has so quickly moved from code generation to code review that it is actually a bit jarring. None of the current systems / norms are setup for this world yet.
-
How Students Use Codex: Key Use Cases
By
–
working on something, if you’re a student – how do students use codex? what use-cases does it support you with?
-
DaaX Achieves Industry-Leading 77.7 Score on FACTS Benchmark
By
–
https://
daax.ai Achieves Industry-Leading Score of 77.7 on FACTS Benchmark: https://
daax.ai/press/facts-be
nchmark
… Technical note on DaaX's methodology and results: https://
daax.ai/blog/facts-tec
hnical-note
… FACTS benchmark: https://
deepmind.google/blog/facts-ben
chmark-suite-systematically-evaluating-the-factuality-of-large-language-models/
… … "The FACTS Benchmark Suite, introduced by Google -
ChatLLM Creates Interactive Charts from Simple Prompts
By
–
Create interactive charts in seconds with ChatLLM by Abacus AI.
— Abacus.AI (@abacusai) 16 mars 2026
Give it a prompt, and it researches the data, organizes it into tables, summarizes key insights, and generates both static and interactive charts you can explore. pic.twitter.com/i5Eg1qlHxeCreate interactive charts in seconds with ChatLLM by Abacus AI. Give it a prompt, and it researches the data, organizes it into tables, summarizes key insights, and generates both static and interactive charts you can explore.
-
Mistral AI joins NVIDIA Nemotron Coalition for open source AI models
By
–
Looking forward to building frontier open source AI models together with @Nvidia as we join the Nemotron Coalition and start training the first base models.
→ View original post on X — @arthurmensch, 2026-03-16 23:24 UTC
-
Evaluating AI Claims: Separating Hype From Real Use Cases
By
–
Try it for yourself and make up your mind. There’s defo some claims that are inflated, lots of real good use cases too!
-
GPT 5.4 Success: Humanity Over Raw Technical Capability
By
–
It is also very smart, but…I generally agree with this, and really feel it myself on the 5.3 -> 5.4 upgrade. rohit (@krishnanrohit) GPT 5.4 is very good, but its most distinguishing characteristic is its humanity. 5.3 Codex was already incredible at coding so it's interesting to see what made it so much more successful. People claim they want a 10x autist savant coder, but what they want is personality. — https://nitter.net/krishnanrohit/status/2033616786978181185#m
-
GPT-5.4 API Launch Achieves Record-Breaking First Week Performance
By
–
Great first week for 5.4 in the API. Builders building fast. Greg Brockman (@gdb) gpt-5.4 has ramped faster than any other model we've launched in the API: within a week of launch, 5T tokens per day, handling more volume than our entire API one year ago, and reaching an annualized run rate of $1B in net-new revenue. it's a good model, try it out! — https://nitter.net/gdb/status/2033605419726483963#m