Loading activity…
@mlstreettalk
MLST is by Dr. Tim Scarfe @ecsquendor w/ cameos from @DoctorDuggar https://www.youtube.com/c/MachineLearningStreetTalk (early access/priv.discord) - Sponsor us!
AI
The reported MATH-500 score rose from 7.9% to 48.1% with the same sampling settings.
AI
Cohere puts the valuation at C$35.7 billion; Mark Carney says Canada is taking an equity stake.
AI
Artificial Analysis ranks its High configuration first on a 22-model index but eighth for live-conversation preference.
AI
SemiAnalysis reports roughly 20% less monthly API-equivalent value but roughly 20% higher five-hour usage limits for some accounts.
AI
Its developer says Flint exceeds 125 tokens per second on an M4 Pro MacBook Pro at short contexts.
XAI
Jesse Singal says a detector relying on Pangram rated QualiaQuanta’s whole article 100% human.
AI
A researcher says an early-reject phase at SaTML cut submissions by about 40%.
AI
Goodfire researcher Nick Cammarata argues that modern tools reveal most neurons as legible in models up to trillions of parameters.
AI
A user says companies they've spoken with need more people to operate AI tools, not fewer.
AI
MLStreetTalk says a post calling it the first evidence of recursive self-improvement drew 1.8 million views.
AI
An October 9 post says 'we're reaching' those scores now, without giving any figures.
AI
A post also highlights why evolved networks find structure that stochastic gradient descent (SGD) misses.
XAI
A post quotes AI as a 'superhuman specialist' that's 'sub-human at running itself,' then speculates about what that gap could lead to.
AI
ML Street Talk shares a video featuring PolyAI's Shawn Wen discussing how the new model handles turn-taking.
XAI
Adaption AI says it released the report while asking how to post-train a model for a new capability with zero data.
AI
Greg Kamradt said $37K would be awarded to the top open-source solutions for Milestone Prize #2. He said Tufa had said it wouldn't open-source this time, after doing so the first time.
AI
A user quoted a claim that “it” would net out at “half the dollar in API spend” compared with the old Pro $200 plan. They praised the transparency but said it “hurts a lot.”
AI
A user says Sonnet 5.5 is better than Astra and has canceled their OpenAI subscriptions. They praise Opus 5.5 but argue that Sonnet’s high output-token use eats into its speed and cost gains.
AI
A user shared a photo link in a reply, then replied to their own post with another, writing, “And another one :).”
AI
The Wall Street Journal says the model was slated for ChatGPT and Codex. CNBC says OpenAI found it fell short of the company’s safety standards.
