Thomas Dijkstra (@thomasdijkstra) header
Thomas Dijkstra (@thomasdijkstra) profile picture

Building & marketing real-life experiences for 200k+ visitors/year. Posts on, AI, code, ads, ops.

Amsterdam, The Netherlands
Joined December 2017
Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)Photo by Thomas Dijkstra (@thomasdijkstra)
i'm ready, thats why im doing real life experiences and not b2b saas
Oh man, the world is super not ready for post-labor economics.
2
225
could making RL environments around this ability and cooking big models lead to RSI, can somebody smart let me know
i've reproduced the results from the MIT blog post today on weight-space programming. but one additional result i have to add: this is a new, step-change emerging capability that's appeared in just the latest set of models.
21
will editing random/fresh weights of a model do anything? like will it actually generate words? did somebody try this ?
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
13
im so fascinated by emergence theory, in life, complex dynamic systems and obiously llms
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
8
GPT 4.5 with a good prompt came the closest of writing good copy in my unscientific tests, although opus 5.5 with fresh/minimal context also pretty good
Crazy how Large Language Models have gotten good at pretty much everything other than generating text
37
clearly did not use sonnet 3.5 when it came out with cursor composer, it was the inflection point.
Ranking all AI models in History
20
I should ship more and let smarter models in the future fix the slop
18
leverage Hoeffding’s inequality in T3 Code: average many independent, bounded measurements and the result almost never lands far from the truth. the odds of a miss shrink exponentially as you add more. “independent” is the important part tho. that’s why my subagents aren’t all the same model. tl;dr: give the same task to several different agents and go with what they agree on
26
something random i learned today: in 1986 a lake in cameroon suddenly burped out over a million tonnes of co2. heavier than air, it rolled down the valleys at night and killed ~1,700 people in their sleep.
30
ram sticks well spend
Obliteration Made with @AskVenice Language: Pali
39
what I love about t3 code is all these quality of life updates, thanks @jullerino @theo
1
90
finally google gave anthropic a reason to drop fable 5.5
🚨 BREAKING: DeepMind have announced Gemini 4 Carbon. Per these benchmarks, it’s the new best model in the world, and leads Artificial Analysis’ Intelligence Index v5
1
68
eu: gdpr cookie banners everywhere also eu: ⤵️
‼️ EU governments approved a GDPR change to let companies train AI on Europeans' personal data without consent. It makes AI training a possible "legitimate interest", judged case by case. Parliament must still agree. Max Schrems calls it a "digital expropriation" of Europeans.
30
👀
Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO. Read more: anthropic.com/news/s1-sec Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.
30
how are the official php docker images still not live lol.
1
234
can the official PHP docker images be build already thank you
1
1
601
hmm need the patched php docker image though
670