eknkc

6 ms·

31.
by eknkc 6mo ago
Not at the moment apparently. They remove the thinking messages when you continue after 1 hour. That was the whole idea of that change. So the LLM gets all your messages, its responses etc but not the thinking parts, why it generated that r
32.
by eknkc 6mo ago
Well anectodally, 5.4 was already better than opus 4.7 so it should not have been hard.
33.
by eknkc 6mo ago
I’ve been using codex cli and GPT 5.4. It is better at coding than Opus anyway. I did not really test Opus 4.7 but older versions generated worse results compared to GPT. Which I would not even try and test though if Anthropic did not ban m
34.
by eknkc 7mo ago
This is a completely baffling website but as far as I can tell, the 2.7 WASM thing is the MCP runtime this is marketing? The globe thing is independent of that, just showing there the MCP calls are running.
35.
by eknkc 7mo ago
And also htmlx with htmx I guess?
36.
by eknkc 7mo ago
I thought I hated React until I saw the samples on this page...
37.
by eknkc 7mo ago
I have a desktop computer that I use for gaming so it had windows forever. Lately it started running laggy. Occasional frame drops and stuff. Reinstall, bios update etc nothing helped. For debugging I installed Bazzite (Linux gaming distro)
38.
by eknkc 7mo ago
At least in codex you can configure agents as you wish: https://developers.openai.com/codex/subagents Might work out fine on codex.
39.
by eknkc 7mo ago
I believe this is the bet in question: https://polymarket.com/event/iran-strikes-israel-on This caught my attention today as it was in dispute and there were a lot of comments (moslty angry). I like browsing polymarket
40.
by eknkc 7mo ago
I find that Opus misses a lot of details in the code base when I want it to design a feature or something. It jumps to a basic solution which is actually good but might affect something elsewhere. GPT 5.4 on codex cli has been much more rel
41.
by eknkc 7mo ago
This kind of stuff can be a great LLM benchmark as Opus basically screwed it up and created a monstrosity as solution on first try.
42.
by eknkc 7mo ago
I had forgotten that mac has a beachball cursor. Something’s wrong on your macbook. (M4 max here)
43.
by eknkc 7mo ago
That is more like a “wish” in windows.
44.
by eknkc 7mo ago
I find time to first token more important then tok/s generally as these models wait an ungodly amount of time before streaming results. It looks like the claims are true based on M5: https://www.macstories.net/stories&#
45.
by eknkc 7mo ago
Not sure. I'm a heavy AI user at this point. Oh, also a heavy Apple user and never once used an Apple AI thing since they released them. I don't even know what they released. It is complete failure of execution on their part.
46.
by eknkc 7mo ago
I still have no idea wht to do with my previous gen iPad Pro..
47.
by eknkc 8mo ago
Everything went down but seems to have recovered now.
48.
by eknkc 8mo ago
I’ve been using OrbStack instead of Docker Desktop and gotta say, I’d not replace it with anything else. So if anyones looking for a more automated alternative, check out OrbStack.
49.
by eknkc 8mo ago
That's my hope but Google has unlimited cash to throw at model development and can basically burn more cash can openai and anthropic combined. Might tip the scale in the long run.
50.
by eknkc 8mo ago
I hope they fail. I honestly do not wish Google to have the best model out there and be forced to use their incomprehensible subscription / billing / project management whatever shit ever again. I don’t know what their stuff cost.
51.
by eknkc 8mo ago
I only put things when the LLM gets something wrong and I need to correct it. Like “no, we create db migrations using this tool” kind of corrections. So far it made them behave correctly in those situations.
52.
by eknkc 8mo ago
A keyboard fix is the only thing I want at this point. It is comically bad.
53.
by eknkc 8mo ago
I predict 100 returns in total and then everybody forgets about this in a week. I'm not cheering for this outcome but it's the sad reality.
54.
by eknkc 8mo ago
They will or they will perform worse. OpenAI will be able to offer a better `free tier` as it won't be free after all.
55.
by eknkc 8mo ago
I've been on Mounjaro for 4 months now. You normally start with an initial lower dose and increase the dose after a month. I lost 5 kg (~11 lbs) on the initial dose in the first month, and when I increased the dose, I also stopped smok
56.
by eknkc 8mo ago
Is the mechanical keyboard craze still going on? At some point everyone was talking about / showing off their mech keyboard in developer scene. I don’t think I’ve seen much in recent years. I myself went deep into that for a while. Got
57.
by eknkc 8mo ago
Yeah but just look at what happened within the last 2 years. I was not convinced about the AI revolution but I bet in another 2 years, we won't be looking at the output..
58.
by eknkc 9mo ago
It is not opus. It is good, works really fast and suprisingly through about its decisions. However I've seen it hallucinate things. Just today I asked for a code review and it flagged a method that can be `static`. The problem is it wa
59.
by eknkc 9mo ago
I've been using it with opencode. You can either use your kimi code subscription (flat fee), moonshot.ai api key (per token) or openrouter to access it. OpenCode works beautifully with the model. Edit: as a side note, I only installed
60.
by eknkc 9mo ago
I've heard people granting access to their production servers to this thing. Apparently you can ask it to check logs to find solutions to some errors or whatever. Gotta be a complete moron to do that. I've only installed it on a f

More ›