
I do stuffs with computer and mentoring programmers. Staff Machine Learning Engineer @unity Vector. Backing and advising: https://xitter.gq/t.co/43LFka55J2
Helsinki, Finland
Joined June 2010
- Tweets72.9K
- Following261
- Followers937
- Likes6.4K
Pinned Tweet
I built Qwenchana.
Qwen 3.8 27b omakase serving setup for 2 x RTX 3090
It is working really well especially with 4-16 concurrent which is ideal for coding agent.
I uses it to run it's own benchmark and fine tuning of the parameter afterward as well.
github.com/vtno/Qwenchana
Joy of programming was never typing out characters at all.
My joy in programming came from solving problem efficiently.
Now I get to do that without even typing!
Tino Thamjarat retweeted
"god when is meta ever going to launch muse voice they announced it forever ago"
muse:
🤖 Made with AI
หลาย ๆ คนพูดเหมือนว่า code ไม่มีลิขสิทธิ์ เวลาโหลด lib มาใช้ไม่เคยอ่าน LICENSE กันเลยสินะ ว่าแบบไหนใช้ได้ไม่ได้ 🥲
เรา ambitious กว่าเดิมได้มาก ๆ ครับ เมื่อก่อนอะไรที่มันถูกบอกว่ามำไม่ได้ ยากเกิน เหนื่อยเกินมันปลดล็อคหมด
ผมที่เป็นคนเขียนโปรแกรมที่เคยต้องบอกว่าทำไม่ได้ เพราะ นั่นนู่นนี่ ผมพูดว่าทำได้มากขึ้นเยอะมากครับ
Tino Thamjarat retweeted
The gap between an average engineer and the best engineers was never quite 10x, but it was meaningful.
AI has made the gap at least 100x.
Tino Thamjarat retweeted
A huge amount of the Anti-AI code sentiment massively overestimates the quality of human code outside of a very small set of open source and high quality company codebases. Human Slop is everywhere and can trivially be improved on by any opus level model.
If you're still thinking AI generates slop in 2026 then it's a skill issue. It's you generating AI slop because you don't know how good software looks like.
And I hate the word software factory because it is misleading.
Our Q3 performance was our strongest ever, with strategic revenue growing 59% year-over-year and Unity delivering our first-ever GAAP profitable quarter since our IPO.
We are seeing the compounding impact of our flywheel strategy: more creators make more games on our platform, and more games fuel the need for more discovery through advertising.
I built this for internal use within Unity's Slack.
This mode of work is insane once you get the hang of it.
waaat
Reuters - 'Gaming software provider Unity Software has held recent talks to be acquired by a consortium of investors including private equity firm Silver Lake Partners and Saudi Arabia’s sovereign wealth fund PIF'
'Valuation of $25 billion was discussed in recent weeks'
reuters.com/world/silver-lak…
Tino Thamjarat retweeted
I'm building a new programming language, Zed (once it's released I'll be posting about it at @ZedLang).
I have been working on this language for over 20 years, but never got it built, because the scope of it is so ambitious. It's an entirely new type of language based on category theory, with core design objectives of precise automatic parallelization up to the Amdahl limit; automatic memory management, with no GC and no need for manual malloc/free; and pervasive use of partial evaluation to increase the power of the language, among other very unique features, such as units (`accel = 5 with units("m s^-2")`) as part of the type system.
I gave Opus 200 pages of design notes that I have been compiling over the last 20 years, and the code I had already written, and asked Opus to compile everything I have created so far into a single cohesive and detailed plan. Then I went through many rounds of cross-review between Claude, ChatGPT and Gemini, having them all do critical reviews, then meta-critique each other's critiques, then collaborate on combining their critiques into an improvement plan, until convergence. I also had the agents cross-check every part of the spec with every other part of the spec (all-pairs comparison) to check for inconsistencies, gaps in the spec, etc. Then I had the agents turn the spec into a long series of tiny incremental steps (a topological ordering of what has to be built before what), as well as the DAG-structuring of dependencies underlying the topological ordering.
I then had Opus build a visualizer that shows me all the tasks in the DAG as my coding agents are working on each task, with realtime update as tasks change status (in progress / completed), which I highly recommend when you ask Opus to do some mammoth complex task, because you can see exactly what it's doing and when.
It is absolutely magic to see this coming together before my eyes, and to see how Opus is handling this massive DAG of complex tasks and their interdependencies.
It is crazy to me that Opus is now simply building what I have dreamed of and wished would exist for over 20 years, but never had enough waking hours to implement! I now have an army working for me, completing what I couldn't due to the bandwidth limitations of having just one brain, and 24 hours in a day.
What a day of miracles we live in.
Tino Thamjarat retweeted
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…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
Neuralink’s first product is called Telepathy.
It enables high-speed communication with computers or humans simply by thinking. Output bandwidth will vastly exceed the fastest human speech or typing in a few years. Read that sentence again.

























