8,782 karma · joined September 26, 2020
I had something similar in https://ccode.kronis.dev/dynamic-proxy.html but since stopped development on the tool, there are also better ones out there.
Made it in Wails cause it was the easiest for the UI. App needs 10 MB of RAM, whereas the WebView2 process needs about 100 MB of RAM.
I’m surprised that Chromium doesn’t have a low memory mode that disables all of the crap your local software doesn’t need. Frankly 20 MB I’d be fine with but this browser tech in its current form really adds up.
I wouldn’t care as much if I had more RAM, but can’t get more than 32 GB in this market in good faith.
Guess I gotta rewrite it in .NET with… something for the UI.
Visual Studio Code, but JetBrains would have been pretty cool, especially when you want something that supports other modes of work (e.g. ADE). Now VSC/Zed fills that role and JetBrains don’t compete with them, now even to upsell to their full IDE for more deep work.
Oh to make things more interesting, Air also has a mobile option too, adding to the list. Odd to couple it with an IDE though since you have various IDEs for various stacks (e.g. IntelliJ IDEA vs WebStorm vs GoLand, unless you just shove a crapload of plugins into IDEA) and agentic tools don't always need all of that.
To be fair though, they're probably scrambling to keep up with the AI craze because it cuts into their margins A LOT and throwing everything they can at the wall. Shame they killed Fleet, could have been a cool basis to compete with Zed (or VS Code, more directly) and integrate Air there or something.
Same, it's lovely software for that mode of work!
I also briefly tried out other ADEs, like:
- Kepler (by the GitKraken folks) - https://www.gitkraken.com/kepler
- Paseo - https://paseo.sh/
Kepler had a really nice issue tracker integration but the UI was very vibey and not quite as to the point as I'd like (who needs graph visualization of what the agents do, just give me tools that work), and Paseo had a really nice UI that I could use to drive both Codex and Claude Code, but some OpenCode agents would sometimes hang and I didn't have tools to really do anything about it.
Maybe I should also give JetBrains Air a look again, if I ever end up subscribing to them again (I just moved to Zed for my code editing): https://www.jetbrains.com/air/
Interesting space, with so many additional tools popping up! That said, the OpenCode TUI was really pleasant too.
Some lawyers are extremely litigious.
This might be daydreaming and wholly unreasonable, but I think that it should go like this: “As a developer/publisher you either provide a product available for purchase for a reasonable price (not larger than release price, adjusted for inflation) OR you are not allowed to take down game preservation efforts.”
Can’t afford to patch the game for vulnerabilities in the code or host servers? That’s okay, the community will do that for you after they reverse engineer it.
But that’s not the world we live in. Personally I think that even projects like SPT-AKI are okay.
Whereas on the RPM side Fedora is pretty good for desktop and something like RockyLinux for your servers (essentially carries on the spirit of the old CentOS).
Then again, I never really needed any of the more fancy modern distros, most of the time I just need good LTS and to run Docker, many might prefer Podman.
Most harnesses now support something like this:
/goal "Work on the current plan until 11 AM in my time zone, provide a summary of all the overnight work then."
2. Previously yes (esp. to try out new models on OpenRouter, like GLM 5.3 or Kimi K3.1 Pro and the various DeepSeek ones, Mistral ones etc.) before also having their subscriptions for a bit. Wrote about them on my blog, seemed like nice alternatives to OpenAI and Ant, but didn't give me as many tokens (and that's with the BS those two are occasionally pulling).
3. Codex and Claude Code, for a while had a custom launcher for Claude Code (ccode) for 3rd party providers, also tried out OpenCode (really nice CLI, but barebones web UI/desktop versions), alongside additional ADEs like Paseo (really pleasant, but sub-agent status updates hanged sometimes) and Kepler (cool ticketing system integration, but the UI wasn't very pleasant), before just going back to the two in the end. Same as with MCPs and other stuff, realized I just don't need most of them, near-SOTA models are good enough to automate long form work and work in parallel well enough, all you need are Git worktrees and a decent OS environment to work in. Still waiting for some dumb agent to wipe my drive though.
4. No, closest was Paseo or Kepler, but those tools would help me more rather than beat the agents into shape. The most I do orchestration wise is demand adversarial reviews (main session often spawns sub-agents for implementation and then has sub-agents with clear context for review in a loop, if issues are found, they're solved, then back to review, until nothing too serious remains), seems like even SOTA models fuck up consistently enough that they benefit from such reviews still. Meanwhile the most I do in regards to work is nudge how to best do exploration and division of work, when I know that something will need a bunch of parallelism or not to pollute the main context with bunches of side quests.
5. No. Maybe I should, but given how often LLMs lack decent judgement when the tasks are underspecified, I don't think I'd benefit a lot. My ALL-CAPS-SWEARING/SESSION metric is still higher than I'd expect from a reliable assistant system, despite agent instructions and memories and custom tools and linters and all that. In a word, even SOTA models still do stupid shit and bog down workable designs with complexities and mechanisms and concerns that nobody fucking needs. Computer use is pretty good though.
> When it comes to shooters though, those games aren't really about competition and decisions, they are about giving the human player an experience. I do think strong AI is overkill in such situations.
> Again, in shooters, there's no decision, no illusion of fair play or equal competition based on making decisions, it's all about the human experience, and so LLM-level AIs aren't needed here at all.
No, playing against humans absolutely sucks to an untold level for me - because more often than not at some point it ends up being about who has better aim or reaction times and I lose like most of those.
Stuff like that makes everything from CS, Valorant, Battlefield, Wardogs, Active Matter, Tarkov, ABI, Insurgency Sandstorm and any other shooter game unplayable. Meanwhile purely skulking around corners is no fun and also sucks, because that still doesn't resolve the imbalance, and that's after over a thousand hours and sometimes with SBMM and all. Teamwork can't really fix that, even when I tried playing a healer in Valorant friendlies would get taken out, I'd get picked off, round over, we lose.
What I really want is AI that is worse at aiming than I am but is more numerous and makes up for it with numbers and utilizing those numbers well with tactics. Games like FEAR and Trepang2 and Incursion Red River evoke that feeling, but could definitely be made even better in that regard! It would be strong not because of sniping me from across the map like a lazily made bot, but rather because there'd be grenades going out in my direction, they'd use cover well and suppress me while some others push from the flank, not just rush me with godlike aim.
That is before you consider playing online being a minefield (especially in Eastern European servers), it's a coin flip whether someone in Wardogs will be trying to flip choppers or AFK bot at base, or whether they'll decide that being annoying and sharing the best slurs they know is what they want to do today. It seems like there's a narrow band between tryhards who berate you for not playing good enough and people who are just trolling. I don't want that kind of hit or miss experience, ever. It's even the same with Arma RP servers (like Altis Life, the idea is cool) when I tried those - you'd be able to make money by mining and processing ores or whatever for a bit, until someone would randomly shoot you for no real reason.
Nowadays: Electron slop that will also eat a bunch of their RAM.
Personally I think that’s due to the failure of native development properly taking off: you know, actually being good and easy instead of a long line of dead or half dead attempts; alongside platform vendors hating each other and each having a vastly different GUI framework instead of some inter-OS WG to come up with at least good common standards.
I mean if you have only one there’s https://www.qubes-os.org/ but not sure about the gaming part, nor is it necessarily as good as separate isolated hardware.
Architecture page: https://doc.qubes-os.org/en/r4.3/developer/system/architectu...
There's horror stories of automated systems suspending people's Google accounts and suddenly they can't access their photos and other important data.
Like if you use a password manager and all of the passwords are random, might as well go a step further and at least have separate e-mail accounts for groups of activities.
The only difference is the architecture/moving parts/featureset which roughly corresponds to the attack surface, which then roughly corresponds to how many CVEs will be found, though that's also affected by how big a given target is.
For example, total new CVEs:
2024 2025 2026
Forgejo 1 10 >55 (confirmed ones, checked what I could)
Gitea 1 9 121
GitLab 145 165 251
That's not to say that Forgejo is some perfect piece of software, just that it builds on some of the past experience of Gitea and just has both less features than GitLab, is simpler, and also a smaller target. It will probably get >100 per year too soon, but I'd reckon that the totals will remain than GitLab for the foreseeable future.
Maybe I'm wrong!
Just get the source code of Forgejo, audit it thoroughly and run an instance of that.
Sometimes how projects are run is weird, like how Hi-Fi Rush seemed to be doing well and then Microsoft axed them despite it:
> In his letter to staffers, Booty said the closures were “not a reflection of the creativity and skill of the talented individuals at these teams or the risks they took to try new things.” And while the changes will be “disruptive,” Booty said that they are “grounded in prioritizing high-impact titles and further investing in Bethesda’s portfolio of blockbuster games and beloved worlds which you have nurtured over many decades.”
https://arstechnica.com/gaming/2024/05/microsoft-shuts-down-...
I guess that’s just the old “Build One To Throw Away” in a modern light, where prototyping more freely is now actually feasible. I think if you apply the good engineering ideas of the past decades to AI, the results can be really good!
Except I’d also add a lot of testing to the prototype, to bring up as many edge cases as possible.
How could that be made to happen? Should EU somehow move to federated (let’s say per country) instances of Mastodon or something? Or one big platform? How would regular folks get used to it? And how could those initiatives be prevented from falling into irrelevance when network effects are at play?
I agree with the idea, it’s just that there’s the usual hurdles there aren’t necessarily good answers for that I know of.
That said, I bet that’s how eventually we’ll end up with actually good ADEs (and a bunch of dead projects), by taking what works and changing or discarding what doesn’t - for example GitKraken Kepler has a really nice ticketing system integration but also the older versions I’ve tried had vibe slopped UI.
If nothing else, competition should push everyone who is capable to keep up to do better. It is kind of a bummer sometimes, though - I also wanted their Fleet to succeed, but that was before Zed became really good.
To be fair, there were some arguments to the opposite, but an odd situation to be in anyways: https://news.ycombinator.com/item?id=49654350
Though to be fair Opus 5.5 is better than Opus 5 was in that regard. I still prefer the OpenAI models to either of them when it comes to regular prose, though.
If they didn't, they'd be unable to compete with their competition and also we'd get worse models that can't reason about many domains or do that much due to limited training data/generalization. I don't think we can put the cat back in the bag or mandate only non-problematic training codebases without shooting ourselves in the foot.
Now, if those AI orgs were mandated to properly compensate everyone whose data they've mined, to track which neurons get invoked and to pay accordingly based on what was used to train them... ehh, that's about as unrealistic as stopping. Kind of a bad situation all around, unless you abandon the idea of any sort of copyright.
It's an excellent way of showing how unreliable the information you find online is (and it will only get worse in the future), but willfully creating misinformation and made up shit and offering it as fact feels hostile to humanity.
Surely nobody got hurt as a consequence of those "facts", but bad decisions with real world impact could be made regardless. It feels like the behavior should be shunned in a general case (the same way I'd shun those numerous spam sites that pollute search results and very much would prefer they not be allowed to exist), even if in this case the intent was good.
Problem is, you can't really do anything about it, other than figuring out a way to take down hostile sites and keep them from popping up, so it's even worse than fraud or even just bad research in academia.
I feel like an increasing amount of orgs out there have long abandoned this thinking.
I wanted something in the EU, so I went with INWX - their UI is quite unintuitive/bad in my eyes though. I guess Hostinger/IONOS/Gandi/... would also be pretty popular for anyone like me.