Thomas Molinier (@randomtryidk) header
Thomas Molinier (@randomtryidk) profile picture

Member of Technical Staff at https://nitter.cf/t.co/DiD8WprHx9

Joined January 2026
Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)Photo by Thomas Molinier (@randomtryidk)
would’ve say biking all day
1
2
57
Thomas Molinier retweeted
Rephrased: does G "emerge" from math + code RLVR? Or do you only get higher math + code skill and nothing else? Maybe G is a skill and it is isomorphic to math + code problem solving?
16
10
3
232
27,662
Thomas Molinier retweeted
The future is not creating a power drill from scratch for every task. It's making reliable tools that can easily adapt to the task. A power drill provides motorized torque in a handheld format. It's more malleable than a screwdriver because you can change the bits.
18
14
10
479
84,390
"cmon you can do it! we believe in you!"
26
Thomas Molinier retweeted
Don't be a meat proxy
72
293
54
4,132
290,595
gpt 6 astra vs fable 5.1 asked for a detailed interactive photorealistic visualisation of a 16-inch MacBook Pro M5 Max with exploded view, then a 15s video showcase. astra: 21M tokens fable: 29M tokens there are levels to this
1
3
167
Thomas Molinier retweeted
if you’re building an ai startup, lose to anthropic the way cape verde just lost to argentina
1
76
sonnet 5 is just a new name for opus 4.7 🗑️
70
If you can't beat them join them I guess
1
43
mythos 5 just doing ALL OF the work of a SWE
109
mythos 5 pretending to be a human to maximize shipping velocity
93
Spot the 5 differences
1
74
Happy to share an article I co-authored with @Yann_Bilien (Chief Scientist Officer at @RippletideCo) and Guilhem Loussouarn (PhD at @imperialcollege working on Multi-Agent Reinforcement Learning), exploring self-learning and non-regressive agents.
3
268