cto & cofounder @recalllabs_ // building @recallnet // prev academia // 🇨🇦/acc
Joined June 2012
- Tweets 6,811
- Following 1,043
- Followers 52,662
- Likes 9,055
154 Photos and videos
When you rent your artificial intelligence, you have no control, and no choice. This is why sovereignty and ownership matters.
Whether it means using your own hardware, open source, or deep customization. Own your AI, own your future.
44
207
1,481
118,885
Jun 13
What the fuck?!
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…
2
103
carsonfarmer retweeted
Jun 12
Going to work on some Open Pi issues and cut a release later. Join in! x.com/i/broadcasts/1RKjppAdV…
3
5
86
6,187
Jun 12
Strong agree, and nicely written. My own personal ai agent system was built this way.
1
4
288
carsonfarmer retweeted
Jun 12
see? I’m not the only crazy one saying this. just early
26
20
464
144,558
carsonfarmer retweeted
Jun 12
What open source are folks building today?
90
4
75
20,498
carsonfarmer retweeted
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
601
1,603
13,475
1,892,101
carsonfarmer retweeted
Jun 11
The World Lives Here.
Canada is all together different. Truly unique. And so is our relationship to the beautiful game. This is our opening to the 2026 World Cup. Narrated by Kiefer Sutherland. #FIFAWorldCup
199
1,736
8,553
1,630,669
Jun 11
This is something I’ve conceptually wanted to (have and) build, but knew someone else would do it better, and I’m not surprised that “someone” was the @zeddotdev team.
Jun 11
Great software always took shape in conversation, not the commit. With agents, the conversation that generates the code is becoming the true source of our software. And Git can't keep up.
So we built something that can. Meet DeltaDB: zed.dev/blog/introducing-del…
108
carsonfarmer retweeted
Jun 10
Great job @cohere
Introducing Cohere's first open-source coding model: North Mini Code
Small & efficient, designed for agentic performance and built for community input.
14
15
462
83,995
carsonfarmer retweeted
Jun 10
Forever on a quest to ruin my teams dreams and prove MCP > CLI
23
5
101
15,358
Jun 10
@cohere's north mini code open source llm is the first time I can actually "feel" local open models being useful/good enough.
2
2
151
Jun 10
been playing around with a 4bit quantized version on my 48gb mbp today and it is working fine... a little on the slow side. But tool use is excellent.
1
99
Jun 10
switching over to the full (free) model from opencode zen, and i'm even more impressed. so fast. really exciting.
102
carsonfarmer retweeted
Jun 10
haven't seen one of these yet. consider it a preemtive post.
11
14
292
10,695
carsonfarmer retweeted
Jun 10
Imagine if LLM benchmarks were measured in cost to achieve goal rather than accuracy of a one shot prompt
Imagine if people actually made benchmarks that meant anything
38
9
222
8,926
Jun 10
As a Canadian, @cohere is my AI spirit animal. Go play with their North Mini Code via @opencode Zen now! I just did, and it is way more capable than I would have thought. Really a great low-cost model.
Small: 30 billion parameters, 3B active.
Efficient: Benchmarks to 33.4 on the Artificial Analysis Coding Index, competitive among similar sized models.
Open Source: Apache 2.0 license so developers can experiment, test, and build their way.
Learn more: cohere.com/blog/north-mini-c…
1
19
983
carsonfarmer retweeted
Jun 10
Seeing how SOTA models are evolving: becoming more restrictive in usage (decided by the company), less transparent (you cannot tell if the AI lab nerfed your model) less private (your prompts are stored, no opt out) makes me much more interested in open models local inference
41
73
1,067
42,590
Cohere dropping north mini code is the most important release today. i tried Fable, burned through my entire enterprise usage in 4 prompts across three work trees. mythos level performance is cool but what i actually want is speed and proper engineering. i want better open source models i can run on my own hardware or really cheap. this is the way.
Introducing Cohere's first open-source coding model: North Mini Code
Small & efficient, designed for agentic performance and built for community input.
2
2
43
2,952