scosman

7 ms·

91.
by scosman 5mo ago
yes exactly. Too many people ask AI to one-shot complex tasks, and wonder it behaves like a junior asked to rush something. I have my own skill: 5 rounds of research/planning/test-planning. Interactive with me in loop for all impo
92.
by scosman 5mo ago
The README has some renders. Some fun ones: 100/100 Acid test without a JS engine, convincing Google SERP even when blocked by bot-detection, makes my homepage look better than in real life.
93.
Show HN: Cursed Browser – a VLM reads the HTML and hallucinates the page(github.com)
7 points by scosman 5mo ago|1 comments
94.
by scosman 5mo ago
If you steal my car, no who knows it's stolen would say it's "yours". We're not talking abstract language concepts, this is a specific case. The data was taken without license/rights/approval. It's st
95.
by scosman 5mo ago
Sure, but the difference here is the pirate is claiming it's "their data" and asking for donations.
96.
Show HN: Videowright – we made our launch video with Claude Code (open source)(github.com)
1 points by scosman 5mo ago|0 comments
97.
How we made our launch video with Claude Code (and open-sourced the tool)(kiln.tech)
3 points by scosman 5mo ago|0 comments
98.
by scosman 5mo ago
We won’t be doing it in 2 years. By then my side will have won!
99.
by scosman 5mo ago
Doesn't work on Chrome or safari for me
100.
by scosman 5mo ago
"Join the livestream" button does nothing?
101.
Git-Backed SaaS: SaaS Without a Service(kiln.tech)
2 points by scosman 5mo ago|0 comments
102.
by scosman 5mo ago
> I am optimistic Kimi K open models soon will outperform Opus models Hard to outperform the model you distill...
103.
by scosman 5mo ago
I believe GDPR says you can access a copy of data they store about you, not that they can't sell it.
104.
by scosman 5mo ago
Did anyone else read this in Anthony Bordain voice?
105.
by scosman 5mo ago
Why so narrowly eval just with/without skill? Same approach is useful for everything: model, params, prompt, sub-agents, skills, rag, etc?
106.
by scosman 5mo ago
Not sure I follow the logic. The list of profiles I visit feels like it’s my data, not the owners of target profile. By that logic can I GDPR chrome for the browsing history of anyone who has visited my site? IANAL but I thought GDPR is abo
107.
Cursed Browser: web rendering engine using visual-LLMs(github.com)
8 points by scosman 5mo ago|1 comments
108.
by scosman 5mo ago
But calling it OpenAlacritty would be worse, which is what happened here.
109.
by scosman 5mo ago
2 commits, 8 hours old....
110.
by scosman 6mo ago
Mythos prompt: Hey Mythos, make me 20,000 H100s.
111.
by scosman 6mo ago
We're officially down to one 9 of uptime over last 90 days: https://status.claude.com
112.
by scosman 6mo ago
I think the thesis is it's not solely about returns. Get a 7% return investing a dollar domestically, plus add a taxable dollar to economy (and some recurring benefit to tax base) beats getting a 9% one-time return investing internatio
113.
by scosman 6mo ago
this is what the unsloth post is for
114.
by scosman 6mo ago
not if I can help it! https://github.com/scosman/pelicans_riding_bicycles
115.
by scosman 6mo ago
Making watches. You can start simple ("modding") which is essentially assembling a movement, case, hands, dial, etc. You can keep going deeper: try to fix or build a movement gears.
116.
by scosman 6mo ago
Agreed. They had a rough patch around the 4.7 to 5 upgrade. New architecture required hardware migration. The 5 to 5.1 upgrade was much smoother (same architecture new weights). As you say, little rough around edges, but still great value.
117.
by scosman 6mo ago
Check out z.ai coder plan. The $27/mo plan is roughly the same usage as the 20x $200 Claude plan. I have both and Claude is a little better, but GLM 5.1 is much better value.
118.
by scosman 6mo ago
also: inference time scaling. Generating more tokens when getting to an answer helps produce better answers. Not all extra tokens help, but optimizing for minimal length when the model was RL'd on task performance seems detrimental.
119.
OutClaw: Manage parallel OpenClaw instances in Docker(github.com)
1 points by scosman 7mo ago|0 comments
120.
Show HN: OpenAPI 2 Skill – progressive disclosure of OpenAPI specs for agents(github.com)
2 points by scosman 7mo ago|0 comments

More ›