KronisLV

7 ms·

91.
by KronisLV 2mo ago
I also self-hosted GitLab for some years, eventually ended up moving over to Gitea + WoodpeckerCI + Nexus (with Nexus for a while). Well there was DroneCI in the middle for a bit there as well, but Woodpecker is perfectly okay and more free
92.
by KronisLV 2mo ago
Don’t ask for permission then? Implement features at a viable pace and drown any unrealistic expectations in reports of CVE fixes - usually a version bump of dependencies, but quite often also some migration work within the app; write tools
93.
by KronisLV 2mo ago
> Where are the amazing digital products/stuff? I've done amounts of refactoring and fixes and written tooling that just wouldn't have happened before. I'm not sure what amazing new stuff y'all expect but the amo
94.
by KronisLV 2mo ago
> difficulty of smuggling the chips in I wonder how long it'll be until they won't need to do that due to domestic options.
95.
by KronisLV 2mo ago
> If you're paying $250k/year for a software engineer, it likely makes sense to have them spend $10k/year Would be nice. I feel like the more you go towards the 3rd world (or tbh anywhere not in the US), the more this calc
96.
by KronisLV 2mo ago
> when you send a feedback report, it attaches the 5 latest sessions to the report with no notice or heads up or anything. Not informing the user of that feels wrong, and even then you'd probably want like a dropdown to choose betwe
97.
by KronisLV 2mo ago
> since neither IBM nor Oracle is a service that any self-respecting person not part of an enterprise sales cycle would actually use Heh, snarky. I guess (almost) everyone uses PgBouncer because they want to use a setup that will need th
98.
by KronisLV 2mo ago
> I suspect it's because the LLM simply does not focus well enough on a single method. Consider asking it to use sub-agents or dynamic workflows for that, then the context will be polluted less by the surrounding stuff, UNLESS you w
99.
by KronisLV 2mo ago
Both this and the parent takes are exceedingly odd - don’t you think that it’d lead to the language features also just getting the same proofs, as well as better libraries and frameworks? And do you not think that this would also lead to th
100.
by KronisLV 2mo ago
What model are you using and in what harness? So I know how to avoid it.
101.
by KronisLV 2mo ago
UUIDv7 for all!
102.
by KronisLV 2mo ago
I'm reminded of the story of Randy, that I watched unfold over quite a few months - he couldn't help himself but to work on new features, rewriting existing features, hacking to make his workflow better, but really struggled to re
103.
by KronisLV 2mo ago
Update: tested it out myself on their Max plan, on some parallel agentic sessions. Currently 20% of my 5 hour limit and 4% of my weekly limit. Total: 58.46M GLM-5.3 Cached: 56.91M GLM-5.3 Uncached: 1.23M GLM-5.3 Output: 315.18K
104.
by KronisLV 2mo ago
Or (somewhat) low TDP GPUs for that matter, like workstation ones, that might have enough total VRAM but not the best bandwidth/compute.
105.
by KronisLV 2mo ago
> ...but no. They do not beat opus on real-world usage. I agree, but then we just need meaningful benchmarks that clearly show that! Otherwise it's hand waving about something that should be put on paper in quantifiable terms.
106.
by KronisLV 2mo ago
I hope really badly that we'll get a new 35B A3B or similar MoE model! I also miss the Qwen 3 Coder Next, which was 80B A3B, there are quite a few use cases where a non-dense model <100B would be the sweet spot (when you have the VR
107.
by KronisLV 2mo ago
My current view on things: Paseo had a really nice UI/UX, except sometimes sub-agents within OpenCode sessions would hang. Still, quite pleasant if you want something like the Codex or Claude Code desktop apps, but across various provi
108.
by KronisLV 2mo ago
Oh yeah, the second one is my table (mostly just formatted what usagisushi started and added some more data points). I am already using GLM 5.3 with their coding plan, but oddly enough the API prices don't really seem to be out yet: h
109.
by KronisLV 2mo ago
I'd say being exposed to the slop tone too much: https://blog.kronis.dev/blog/ai-slop-is-a-self-inflicted-tra... because the technical capabilities of the model and the harness (at least nowadays) are all fine. Th
110.
by KronisLV 2mo ago
What the person above is suggesting: * https://pi.dev/ * https://omp.sh/ (no personal opinions of either, links might be useful) I think that OpenCode is nice, their CLI version is enjoyable and their deskt
111.
by KronisLV 2mo ago
I am not talking about GLM models being served through an Anthropic compatible API, that part is perfectly fine and I'm glad they support it! I am talking about ZCode, the program, being unable to delegate to other harnesses, like usin
112.
by KronisLV 2mo ago
Their coding plan switched to credits, didn’t it? What are the rate limits like, compared to Anthropic or Kimi K3? I remember trying their Coding Plan out before the change and the 5 hour limits felt too restrictive then even for light&#x2F
113.
by KronisLV 2mo ago
Yeah, seems the same for me! Except the subscription plans you can get like Kimi K3 or GLM Coding Plan are still giving you more value if you need that many tokens (and can use those plans).
114.
by KronisLV 2mo ago
Everything is a plugin? I'm reminded of Eclipse!
115.
by KronisLV 2mo ago
My apologies to any mobile users, but for the desktop folk: Provider, Model Billing Input Output Cache read Cache write DeepSeek
116.
by KronisLV 2mo ago
> The app supervises many agents, with different project views and built-in git worktrees, so agents can work on isolated copies of the same repo. It's easy to jump between tasks and see changes. This is mostly true with Codex and C
117.
by KronisLV 2mo ago
Would be cool to have a single-user ADE tool like that from the Zed folks - because Kepler focuses a bit too much on the ticketing integration and Paseo sometimes has issues with sub-agents.
118.
by KronisLV 2mo ago
> But once you start storing data there, you can't just remove them. You have to come up with a migration plan, make sure you don't disrupt the system because people are paying to use this every day. You have to think about wha
119.
by KronisLV 2mo ago
> the harness has almost equal, if not more weight than the model itself This feels like a horrible failing of the models to generalize, then - both basic and intermediate tasks should be possible to do with Claude Code, OpenCode, Pi, ZC
120.
by KronisLV 2mo ago
> You can run full-fat DeepSeek locally for (just) under $10K USD.) Is that price not way off if you want actual decent performance, like at least 30-60 tokens per second and at least >256k context size?

More ›