Thinking Machines TML-Small 64.7%, MIT Brain Study 🧠, Rust Browser πŸš€

AlphaSignalΒ·Β·6 min read
AI/MLTechnologyScience
Share𝕏in

AI Summary

Thinking Machines released TML-Interaction-Small, a 276B parameter real-time AI model that simultaneously listens, speaks, and processes video in 200ms chunks, scoring 64.7% on timed speech benchmarks vs GPT-Realtime-2's 4.3%. An MIT study using EEG headsets on 54 participants found ChatGPT users showed up to 55% reduced brain connectivity and 83% couldn't quote from essays they just wrote, coining the term 'cognitive debt.' Additional signals include an open-source Rust browser loading pages in 85ms, a new uncensored local video model generating 24fps clips, and Meta FAIR's byte-level model cutting LLM decoding steps in half.

Key Facts

βœ“Thinking Machines' TML-Interaction-Small is a 276B parameter model that simultaneously listens, speaks, and watches video in 200ms chunks, scoring 64.7% vs GPT-Realtime-2's 4.3% on timed speech benchmarks.
βœ“An MIT EEG study of 54 participants found ChatGPT users showed up to 55% reduced brain connectivity and 83% couldn't quote from essays they just wrote, introducing the concept of 'cognitive debt'.
βœ“An open-source Rust browser loads pages in 85ms using 10x less RAM than Chrome, and uncensored local video model Sulphur 2 generates 10-second 24fps clips with no content filters.

Author Takes

BearishAlphaSignal

AI and human cognition

We're building smarter AI interfaces while quietly accumulating 'cognitive debt' β€” the irony is that better AI conversation tools may be making humans worse at thinking.

NeutralAlphaSignal

Optimal AI usage strategy

AI should be used as a finishing tool, not a starting one β€” engage in self-driven cognitive effort first before leveraging AI assistance.

More from AlphaSignal

πŸ“°Today⚑FeedπŸ“‘SignalsπŸ’°Capital