traviscline Profile Banner
tmc/acc Profile
tmc/acc

@traviscline

Followers
10K
Following
16K
Media
284
Statuses
5K

Hacking on cgpt, langchaingo, nlm, OS AI stuff. 💼 ai. prev Facebook, Alto, Virta, Brev, NVIDIA

San Francisco, CA
Joined May 2008
Don't wanna be here? Send us removal request.
@traviscline
tmc/acc
7 months
Not your trajectories, not your tool.
6
0
16
@traviscline
tmc/acc
6 days
Great demos and meetup by @ExaAILabs and @ollama - enabling web search in the open ecosystem is a key step.
0
0
6
@syhw
Gabriel Synnaeve
15 days
(🧵) Today, we release Meta Code World Model (CWM), a 32-billion-parameter dense LLM that enables novel research on improving code generation through agentic reasoning and planning with world models. https://t.co/BJSUCh2vtg
59
306
2K
@traviscline
tmc/acc
16 days
HN discussion here: https://t.co/yb30cxfpAU
0
0
0
@traviscline
tmc/acc
16 days
macOS 26 / iOS users - are you seeing these trustd failures in logs?
@nathanborror
Nathan Borror
16 days
Should we be concerned about this? Seems like a problem on macOS 26 as well.
1
0
1
@traviscline
tmc/acc
17 days
Are others seeing odd trustd failures after updating to iOS/MacOS 26?
0
0
1
@traviscline
tmc/acc
17 days
Strongest quake I’ve felt! Wild.
0
0
1
@traviscline
tmc/acc
1 month
Love the name :)
@mattt
Mattt
1 month
Apple's JS-rendered docs leave AI tools staring at blank pages. I got tired of Claude writing broken SwiftUI code, so I built https://t.co/pb1UMcV8mC Swap developer․apple․com → sosumi․ai in URLs to get clean, AI-readable Markdown. Or connect via MCP.
1
0
0
@traviscline
tmc/acc
2 months
<disclaimer>I can’t share internal chain-of-thought. Here is a concise high-level plan and assumptions.</disclaimer>
0
0
1
@traviscline
tmc/acc
2 months
New model seems apt to refuse sharing metacognitive state.
1
0
1
@traviscline
tmc/acc
2 months
In-api grammar specification 🤤
0
0
2
@traviscline
tmc/acc
2 months
Time is a flat circle.
1
0
1
@traviscline
tmc/acc
2 months
New fine tuning target.
@AnthropicAI
Anthropic
2 months
New Anthropic research: Persona vectors. Language models sometimes go haywire and slip into weird and unsettling personas. Why? In a new paper, we find “persona vectors"—neural activity patterns controlling traits like evil, sycophancy, or hallucination.
0
0
4
@traviscline
tmc/acc
3 months
My last (probably final) hurrah into politics was to support @AndrewYang because he was warning about how we need to prepare for exactly this type of issue. We need more forward thinking folks in office (at all levels).
@vitrupo
vitrupo
3 months
Sam Altman says it's insane that banks still accept voice prints for authentication. AI has already broken most of the systems we use to prove who we are, and we're sleepwalking into a synthetic fraud crisis. “Right now it's a voice call. Soon it's gonna be video. It'll be
0
0
1
@traviscline
tmc/acc
3 months
https://t.co/l86fGUMGiT Yannick on energy based transformers paper
0
0
0
@traviscline
tmc/acc
3 months
0
0
2
@traviscline
tmc/acc
3 months
1
0
1