Joined October 2011
- Tweets 6,641
- Following 1,250
- Followers 2,574
- Likes 6,347
585 Photos and videos
Austin retweeted
Me: "C'mon, it's not true that all Bitcoiners are autistic!"
Bitcoiners:
6
4
51
3,071
Austin retweeted
22h
We don't know who needs to hear this but we have local inference built in
Private by Default: Built-in Local Inference Models with goose share.google/e09zAzYATaDmLgL…
1
6
23
2,189
Jun 13
Time is running out
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
3
160
Jun 12
Bitcoin has decades to win but we only have a few years to win with local open source AI
4
26
645
Jun 12
This is the way
About $1,000 a day the moment.
I am writing 5 projects at the same time, not sleeping and my family all hates me.
1
1
314
Jun 12
Insane!
Jun 12
DiffusionGemma can now run at 2000 tokens/sec! ⚡
We made local DiffusionGemma inference 1.8× faster.
Run it on 18GB RAM via Unsloth Studio.
GitHub: github.com/unslothai/unsloth
Guide: unsloth.ai/docs/models/diffu…
1
4
989
Jun 12
China continues winning in open weight models
Jun 12
🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced!
🔷 Improved coding & agent performance over K2.6: 21.8% on Kimi Code Bench v2, 11.0% on Program Bench, and 31.5% on MLS Bench Lite.
🔷 Reasoning efficiency: Less overthinking, with 30% lower reasoning-token usage compared to K2.6.
🔷 Long-horizon coding: Improved instruction following, higher end-to-end coding task success rates.
⚡️ 6x High-Speed Mode coming soon!
🔌 Available today via Kimi API and Kimi Code.
🔗 Kimi Code: kimi.com/code
🔗 API: platform.moonshot.ai
4
165
Jun 11
Can we get these for the next hackathon?
Coors Light is rolling out a 'Tallerboy' for the World Cup that holds THREE beers in it
On sale tomorrow on Coors website
1
2
261
Jun 10
prompting in a loop (from what I can tell) means one of:
- openai /goal
- anthropic ultracode
- droid missions
- ralph loop
- boutique agent orchestration pipeline
?
Jun 7
Here’s your monthly reminder that you shouldn’t be prompting coding agents anymore.
You should be designing loops that prompt your agents.
1
246
Jun 10
Factory Droid is the only place I am using anthropic models nowadays. Amazing harness but the token plan and desktop app experiences still aren't the greatest
1
5
335
Jun 10
You aren't vibe coding your taxes anon??
Jun 9
Hermes desktop local qwen 27b browser use pulling receipts and tax docs for me
Trying to test the limits of browser control with small local models
1
5
247
Jun 9
Something about the fable of the devouring overprotective mother x.com/rob1ham/status/2064492…
59
Austin retweeted
Jun 9
I love sota models but I’m building localmaxxing.com and having fun
I don’t really need mythos or even gpt5.5
That’s why models like qwen 27b are more important, they push the envelope for everyone, small local set ups, entry level rigs, cheap api access, and decent performance.
With a 2k setup you get 27b models running at 70 tps with infinite token burn
Everyone buys a setup for coding anyways why not make it think for free
9
4
72
4,717
Jun 9
Hermes desktop local qwen 27b browser use pulling receipts and tax docs for me
Trying to test the limits of browser control with small local models
2
1
10
1,534