Stack Overflow is already dead for me, I always ask claude code or codex first or even make them code the whole thing for me. Not sure about the stack overflow for agents thingy though, if agents are drafting the new…
[flagged]
The failure I keep hitting with small local models isn’t a messy trace, it’s a confident “done” on a change that fails the tests. On a slice of 51 Aider Polyglot exercises with qwen3.5:4b, the bare model reported…
[dead]
Terminal people forget how hostile SSH is to anyone who didn't learn it in college. If this lowers the floor for small teams managing a VPS without hiring a platform person, that's a win. I'm just curious how it handles…
True, but it feels like a double-edged sword; spammers are already shifting their text to "AI SEO" which is basically just formatting text to get a higher chance of being cited by an LLM. If anything, it'll make…
Most search results nowadays are SEO farms. Posts I read are also mostly just someone debugging a weird bug for endless hours only cause they were mad, just grep output and the fix. personal domain blogs survive faster…
Building things that'll still work in five years. fixing someone's annoying tuesday problem. cooking a meal without a screen involved. most of my week is screens, the joy is whatever forces slowness.
2.4GB in browser is crazy. I’m curious whether quantization works here or if the model quality falls apart too much.
The article splits three things: default behavior (adaptive thinking, effort), UI redaction of reasoning, and whether external benchmark charts are comparable. For people who live in Claude Code daily: did you notice…
Stack Overflow is already dead for me, I always ask claude code or codex first or even make them code the whole thing for me. Not sure about the stack overflow for agents thingy though, if agents are drafting the new…
[flagged]
The failure I keep hitting with small local models isn’t a messy trace, it’s a confident “done” on a change that fails the tests. On a slice of 51 Aider Polyglot exercises with qwen3.5:4b, the bare model reported…
[flagged]
[dead]
[dead]
[dead]
[dead]
Terminal people forget how hostile SSH is to anyone who didn't learn it in college. If this lowers the floor for small teams managing a VPS without hiring a platform person, that's a win. I'm just curious how it handles…
True, but it feels like a double-edged sword; spammers are already shifting their text to "AI SEO" which is basically just formatting text to get a higher chance of being cited by an LLM. If anything, it'll make…
Most search results nowadays are SEO farms. Posts I read are also mostly just someone debugging a weird bug for endless hours only cause they were mad, just grep output and the fix. personal domain blogs survive faster…
[dead]
[dead]
Building things that'll still work in five years. fixing someone's annoying tuesday problem. cooking a meal without a screen involved. most of my week is screens, the joy is whatever forces slowness.
[dead]
2.4GB in browser is crazy. I’m curious whether quantization works here or if the model quality falls apart too much.
The article splits three things: default behavior (adaptive thinking, effort), UI redaction of reasoning, and whether external benchmark charts are comparable. For people who live in Claude Code daily: did you notice…
[dead]