He said "10X pretraining compute" which doesn't mean 10x bigger. It can well be the same size but use more training tokens, longer contexts, and other algorithmic changes.
RESEARCH
-
Inferring Model Size from Speed and Pricing is Complex
By
–
I think model size is hard to infer from speed & pricing: it depends on – GPU type
– precision and quantization at inference
– batch scaling
– subsidized or not
– MoE or not
– etc. -

Governments Must Invest in AI Understanding and Ecosystem Support
By
–
"Global governments need to invest in the time, effort, and people to understand what this technology is, and support that ecosystem." HAI Co-Director @drfeifei and other faculty and fellows share their insights from the #AIActionSummit in Paris: https://
hai.stanford.edu/news/ai-action
-summit-paris-highlights-shifting-policy-landscape
… -
When does the AI-as-intern analogy fail us?
By
–
We keep saying AI is like a very smart intern: give it more background & context, show it good examples, review its work, assume it has low expertise, assume it won’t be perfect. When is this analogy shooting us in the foot?
-

MIT Building 20 Replaced by Stata Center: 40 Years of Evolution
By
–
40 years ago vs. now: MIT's Building 20 (foreground) & its replacement, the Stata Center.
-

GPT4.5 Vibe Test: Higher Emotional Intelligence and Creativity
By
–
Vibe Testing GPT4.5 OpenAI just released GPT4.5 w "higher emotional intelligence and creativity" than prior models. Here, we overview its capabilities and vibe test it on research / writing using our Open Deep Researcher. Video: https://
youtu.be/Jo2-fZ2THzU -
ChatLLM Research Tool: URL Scraping and Data Analysis Capabilities
By
–
Conduct your research with ChatLLM!
— Abacus.AI (@abacusai) 28 février 2025
It can scrape URLs, extract key insights, and perform analysis. While its depth is still evolving, we're constantly improving it for more in-depth research capabilities. pic.twitter.com/9MUU2txP93Conduct your research with ChatLLM! It can scrape URLs, extract key insights, and perform analysis. While its depth is still evolving, we're constantly improving it for more in-depth research capabilities.
-

GPT-4.5 trained with new supervision techniques according to system card
By
–
Ok, according to the system card (
https://
cdn.openai.com/gpt-4-5-system
-card.pdf
…), it's trained "using new supervision techniques" -
RL and inference-compute scaling potential for model improvement
By
–
I wouldn't discount it quite, yet. I am curious how it'll do with some additional RL + inference-compute scaling à la o1.
-
Missing Technical Report Reduces Excitement About New Release
By
–
I agree, but at the same time I think people (or at least I) would have been much more excited about this release if it came with a Llama- or DeepSeek-style technical report.
It's hard to get excited about something if you a) don't know how it works or b) it's not substantially