Ariel Ekgren (@ArYoMo) header
Ariel Ekgren (@ArYoMo) profile picture

Building @hybridity_ai ex LLM researcher

Stockholm, Sweden
Joined April 2013
Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)Photo by Ariel Ekgren (@ArYoMo)
Ariel Ekgren retweeted
I’ve worked that heatshield my entire adult life. Peak career moment today To go from little 7 tile patches to a full reusable ship. Absolute marathon I’m so grateful I got to do something useful, that I was given the chance I don’t really have many good memories from this experience. It was so difficult. It almost looks easy now, but none of it came easily Please do something amazing with this rocket. This tech is such a gift, a love letter to the species
1,111
2,573
172
31,536
562,792
If the results holds up... Impressed. Why train when we can just let an LLM manually write the weights for the best results.
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.…
1
151
Error starting chat Cannot serialize value: (couldn't stringify value) github.com/openai/codex/issu… @thsottiaux I think you really want to fix this. Might actually leave a mark in usage.
2
1
7
255
Impressive! Meta is really in.
Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.
2
140
I urge you to urge me to stop right now!
Dario Amodei is at the desk to assure that the future of humanity is safe from AI
3
260
Ariel Ekgren retweeted
The constant decel messaging in Europe is insufferable as an adult, but its effect on kids is worse. As a parent, it's your obligation to counterprogram this nonsense. Show them robots, rockets, and wonders of AI. Show them cathedrals and moon landings. Instill hope in progress.
203
704
74
8,920
285,997
Progress we take for granted and hoooly opus 5.5 goes hard
holy shit i asked claude to make a video on western civiization
3
292
I am again impressed. Opus 5.5 made this for me. A glimpse into base reality.
1
115
Ariel Ekgren retweeted
Replying to @francoisfleuret
It's all computation. Math is just a game we play to reason about some simpler parts of computation.
7
4
4
66
63,459
Are we so back?
everyone so busy yapping about AI no one noticed room-temperature superconductor claims are back arxiv.org/abs/2609.15712
1
172
I don't know if it's my brain rotting in real time or if it's the models. I no longer understand what codex and Astra is saying to me.
1
136
Love to see this and hope that it reaches mass adoption sooner rather than later.
Through the VOICE trial, Terry is using his Neuralink implant to help fine-tune a brain-to-voice interface for himself and others who can’t speak. He trained the algorithm first by miming speech as best he could, then by simply thinking the words and hearing them come out in his own natural voice. Powered by Grok Voice from @spacexai
1
216
AI power games
99
We will not monitor chains of thought for long as a way to control the models and that is totally fine.
1
1
152
Do not fall for the AI propaganda. The battle is for power and it's political. There will always be power struggles but it is easier to align with your preferred side if you realize what is going on. AI is on track for amazing feats, not yet autonomously.
1
176
The level of automation you can do already is mind blowing. Limit is imagination, token spend and will power. Not actual model capability. Was hesitating to do some non coding admin because boring & time consuming. Codex just did it, turns out my willpower was limiting factor.
2
275
🙏
man ignore the benchmarks and the race i just want there to always remain a frontier level intelligence model that is open and safe, available to everyone regardless of region it is truly incredible to have built such an open (and strong) model with a group of genuine people
188
Open source is actually at the forefront. Chinese labs very impressive right now.
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3
1
130