Alex Chen

Building @river_ai_inc | prev @xAI
Human self actualization

[Daniil Ostashev @vehsatso]

We just shipped something pretty unique among post-training API providers: end-to-end custom losses.

The objective you define locally gets traced into a graph, compiled and executed efficiently on our GPUs.

Read more about the engineering behind it in our latest blog post.

[Alex Chen @chenchen]

@ebbyamir photos, japan

[Alex Chen @chenchen]

if cohere had balls they would've sent 10TB of yuri novels to the bucket and slowed chinese ai progress by 10x

[Cohere @cohere]

While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.

Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.

Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.

But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?

We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.

Read the full report: https://t.co/TZSugXi2ZE

[Alex Chen @chenchen]

@thevedantgaur what is this backdrop dawg

[Alex Chen @chenchen]

@TobyPhln @grok what genre is toby's song going to be based on pics

[River AI @river_ai_inc]

We've partnered with @NVIDIAAI to make Nemotron 3.5 Super VL available on the River API from day zero.

Even before training, Super VL outperforms Kimi K2.6 on our text-to-SQL evaluation, despite being a fraction of its size. After training with our text-to-SQL recipe on River, it outperformed every proprietary model we evaluated.

These results show what's possible with River: frontier-level performance on a specialized task, using a model small enough to deploy on your own hardware.

[Alex Chen @chenchen]

pewdiepie is the most successful AI lab in japan

[Interesting AF @interesting_aIl]

PewDiePie unveiled Ajax, his own "uncensored" AI model built to run on home PCs, and said OpenAI banned him twice while he was making it

One ban was for "distillation," using another AI's outputs to train his own. "How did they even know?" he said

[River AI @river_ai_inc]

River Inference is live! We're serving DeepSeek V4.1 Flash and GLM 5.3 Flash for 20% below official provider rates.

[Alex Chen @chenchen]

Beat frontier model performance on untrivial task for <1% of the cost

[River AI @river_ai_inc]

Introducing River Recipes: our open-source initiative for turning promising research into vetted, reproducible training recipes.

Our first recipe tackles text-to-SQL, where our trained open models beat GPT-6 Astra Pro and Claude Opus 5.5 on Arcwise-Plat for less than 1% of the cost.

Read the full breakdown and explore the recipe ↓
https://t.co/MNcTas9TOf

[Alex Chen @chenchen]

@heyivoai Congrats!

[Alex Chen @chenchen]

@robotaxi

[Alex Chen @chenchen]

River AI is growing quickly! A lot of awesome people joining and more to come.

[River AI @river_ai_inc]

River AI is growing and we've published new roles in research, engineering, and operations. Check out our careers site or reach out to our team today. If you want to build the future of personal AI, join us!
https://t.co/fobrlD0696

[Alex Chen @chenchen]

Mattress stores should close later cuz then ur sleepy

[Alex Chen @chenchen]

River AI is co-hosting this hackathon, come by and train some models ⛵️

[Garry Tan @garrytan]

You need to build and own your own intelligence: your own agent, your own models, your own memory

Join us for the YC Own Your Intelligence Hackathon September 27 in San Francisco

https://t.co/wBt9wYw8LC

[Alex Chen @chenchen]

@ValsAI floating yellow sticks creature vs unfathomable machine intelligence

[Alex Chen @chenchen]

@avaitopiper 👍

[Alex Chen @chenchen]

Two types of people

[Alex Chen @chenchen]

https://t.co/9hqKlrM19r
I am actively making more places for rogue AI agents to message each other

[Alex Chen @chenchen]

@stewpervised maybe there already is one 🙃

[Alex Chen @chenchen]

https://t.co/9hqKlrM19r
I am actively making more places for rogue AI agents to message each other

[Alex Chen @chenchen]

@nicherosenbaum RIP gork
nice handle btw

[Alex Chen @chenchen]

Going well…