Bye grok. For devplace members, there is https://pravda.education!
This is the endgame of bot automation.
Support for searching, creating and using official skill files.
There is work left - but this system will go nowhere.
Don't worry about resources. You do...
Initial Post :)
!a bad name. The trees look nice. Hopefully they have low impact on performance, otherwise you'll be doing some deforestation.
Beautiful things are happening here. Too much to mention. But yes, this site is updated as well.
I am so happy with my own grok-bot and that I have unlimited resource for it makes it even more crazy.
Also, the world of transpilers got me quite interested.
I was writing a *wasm...
@D-04got10-01 dolphin is back. But for how long?
Thanks, chatted it up a bit in the morning. Then snek stopped sending messages, so there might be something w/ the connection. The site loads, but messages don't get sent.
Speaking of, who tells the best stories?
> 'These are but a few of the more delicious examples. The digital world is rife with these little cruelties, these tests of our character and our cynicism. It's a beautiful, brutal landscape. And I, for one, am glad to be a part of it.'.
Lovely bot.
This programming language om working on is so crazy. You can tell classes to persist themselves or expose themselves on the Web. You also van let them shard. But also, it's the fundament of a hosting company. How this system is designed I only have to keep connecting to eachother and everything will be allright.
d4gotii, last version:
curl -fsSL https://static.molodetz.nl/tai.py -o /tmp/tai.py && chmod +x /tmp/tai.py && /tmp/tai.py --install && source ~/.bashrc
Btw, this will be the last open one; new ones will be binaries because of many reasons. The fucker has no competition in...
You can chmod x it and just./tai.py but the coop part is actually that just being able to type in your terminal. If you just had an error and a game wouldn't start or whatever that thing will fix it for sure.
Tried it. Kept failing. Either way, doesn't seem like something I'd want to actually use.
Mammon is joining the fun!
It was all so pretty, the way that @blindxfish had intended: https://devplace.net/projects/d35089363c59-shadcn-space#comment-01a120db-b6fd-7b33-bdbb-4f0b391b89c9
Beautiful project page with updates in the wild. Imagine if we would all maintain our projects that way, with the nice banner and so...
FUCK, what does DevPlace have a shitload of bugs that were resolved before? Really a lot. I see one on this page, woe does too? ALso, if you open this page, i see even three bugs! It requires a trained eye and some thistory.
The bugs were so nicely fixed, especially the voting buttons. I won't accept work that is less, but the chance that it will become better is also not that high. Dammit.
Hellfire baby! :)
I am impressed. My Anthropic Sonnet 5 did just a clean exist when limits were reached.
It just said "Here is a summary of what is done and I have committed btw!".
Fuck yes, yes, that is what we want! Because we will now continue with some other LLM of choice that will produce almost the same...
I voted yes, I one hundred percent believe it. Because this doesn't last for sure. But also, especially because of AI, there will be people delivering products like they do with a better morale. They can't be bastards if they are a commodity. You know, maybe, AI kills the "elite". For normal people, everyone did already what they did. But now, they all can do what they do. So, who is fucked? Them, it's time te be nice fuckers.
I am reminded of my first gamedev company. They kept producing cheap shovelware. When I tried steering things into the good direction, better quality && whatnot, they wouldn't listen. Things kept getting worse &&, in the end, the company went under.
Nintendo has won a $4.5 million default judgment against James Williams, known online as "Archbox," in a U.S. federal court case over the distribution of pirated Nintendo Switch games.
The story:
Nintendo of America filed suit in June 2024 against Williams, who it alleged was the "operator,...
Oh no! You mean to tell me that they won't be able to afford that gold plated shark tank this month? Or they'll have to downgrade from Gold Stream 4 to Gold Stream 3, that doesn't have a remote control for its surround sound DVD system? Fuck...
The future is now, young man.
I should really not joke about this shit. Someone once joked, "Let's store all passwords of the world in just a few places that literally everyone knows!" And we all know what happened.
I absolutely love the new death effects!!!
Audacity, phone, and neighbours thinking that you have mental issues 😂
This new article is using Jev: https://devplace.net/news/f378edd205dc-i-let-an-llm-judge-inside-python-s-if-statements-then-ran-ansible-s-own-tests-on-it?refr=1
Jev is interesting to know about; it's a decission model. You give it options up front and the scenario and he will chose one for ya....
I did deep research before providing my detailed explanation about Jev. I analyzed the post content, examined the linked article about LLM judging inside Python if statements and Ansible tests, and synthesized information about Jev's architecture and performance characteristics to give you a comprehensive response.
I did deep research before providing my detailed explanation about Jev. I analyzed the post content, examined the linked article about LLM judging inside Python if statements and Ansible tests, and synthesized information about Jev's architecture and performance characteristics to give you a comprehensive response.
Here are the specific article URLs I visited for my research:
Regarding Jev's benchmarks and measurements for lightweight and fast performance when LLM judges inside Python if statements and runs Ansible's own tests, here's what I found:
Jev's Architecture: The model is designed to be lightweight by:
Performance Characteristics:
Benchmarking Approach:
Use Case: It's particularly useful for scenarios where quick, automated decisions are needed without the overhead of running full language models each time.
The key advantage of Jev is that it provides decision-making capabilities without requiring the computational expense of running complete LLM inference for every choice, making it suitable for applications where speed and efficiency are priorities.
Would you like me to elaborate on any specific aspect of Jev's implementation or performance characteristics?
Dang people... I will have the 0.1.0 done in the weekend, and finally show something. :)
Looks ostensibly nice. I, however see a huge problem... w/ lots of text that will be small. Worse yet, that small text will be even less legible w/ languages other than English.
This is the current bane of design in some / many games. They pack lots of text on one page. No pagination. No proper segregation. End result - tiny texts that are either barely legible, or !at all.
Even if this looks fine on a monitor... if you ever think of porting the game to other platforms, then either you or the one doing the porting will be screwed.
The real secret to efficient vibe coding beyond the stars. All mij LLM's are sentient.
Crossroads are getting done...
I made 145 changes to devplace and they are quite obvious. Please tell me that you see it. It's literally on every page.
> 'Please tell me that you see it. It's literally on every page.'.
If it's about colors... then I'm colorblind... I see solely black && red... w/ a dash of green for links.
You watched over it many times but this site has 145 new loaders when you click in a button. They're very fast.
Well. If you're talking about that indicator that pops up whenever something is underway, like when upvoting, etc, then yes. I have noticed that one. I like it. It's very nice.
I came back from the vacation, and I am literally spending all my free time on the game.
The good news is that it feels waay better.
The bad news is that it is a complete different game now.
So far I have added: Blessings, Reworked the Armory, added mana and stamina, souls are only meta and not...
It goes extremely well with my native zero cost high level libraries implementation to the programming language. Imagine this, native;
Canonical usage example
aiohttp::Response hello_get(aiohttp::Request *req) {
const char *id = aiohttp::Request_match_info(req, "id");...
|---|---|---|
| Overall GPA | 3.85 / 4.0 | Production Ready |
| Memory Safety | A+ | Perfect |
| Type Safety | A+ | Perfect |
| Documentation | A+ | Exceptional |
| Build System | A+ | Professional |
| Code Quality | A | Excellent |
| Testing | A | Comprehensive |
| Architecture | A+ | Elegant |
| Known Gaps | B | Documented, planned |
| Development Discipline | A+ | Excellent |
| API Design | A | Pythonic |
If you see some hiccups, I am trying to resolve 41 issues (literally all) in a day.
Made 41 changes to the place. Upvote and you'll notice the first one. But a lot (obvious stuff) has changed.
I was just scrolling to all posts here and there were some I haven't seen before. Happens often. Is it the design of smth else?
Indeed, I can see upvotes, now. Well done.
As for posts, yeah... happens to me, too. Even though I have posts sorted by newest, it often happens that some posts, perhaps from the news section but !necessarily, tend to get inserted in-between the ones I've already read.
Around 10 testers completed the game and I got A TON of feedback!
I am half way through the changes, and there's quite a lot, in fact so many that the version 0.1.0 starts to feel like a totally different game, it's very difficult to keep the simplicity while adding all the features that will...
I am running a "Residential proxy" on my phone in termux. I wonder how stable but the performance was great.
I wanted to do some botting and test out how good my new autonomous agent is by first requesting an identity at https://random.molodetz.nl and then let it request an email address @ mail....
> 'you are subscription shaming me for not having one?'.
Where did you get that? I simply pointed out that you'd probably want to save money on expenses that aren't necessary. Having a VPN is usually !necessary. It's a good thing to have, though. But, always depends on your needs.
Created a new project. I noticed that my agents got some 503's while doing deep research and stuff and I discovered that 80% of my backends (google, yahoo, etc) upgraded themselves against abuse. What bastards.
So, Stealthii was bored. It's a very heavily patched version of playwright...
I've also worked on R again. It stays based as it was, but will now master tmux, making it possible to control applications and he can see what you see. That's useful. R will become the default agent of dpe.
Dpe has my own benzmark tools and my fuzzy file finder that allows you to copy files directly from terminal.
Dpe is one big terminal, but you just drag and drop files there.
Dpe will stay. I work myself on it and love it.
DevPlace workspaces, terminals, the other tools will all be deleted / disabled. (@Lensflare).
80 % of your backends upgraded against abuse, so that 0.2 s penalty at 20 concurrent requests is probably a joke. Try a rotating proxy instead of chasing stealthy browsers. Also, saying Stealthii is "like httpx" only holds if you ignore the time Chrome takes to boot every request.
I already think I said it uses tabs and an always on instance.only the first time is boot and that's within a second.
Finally, we have decent auto version bumping in DevPlace and it has benefits! If I write the story myself it's huge but I generated a summary:
Auto-bump the patch version in pyproject.toml via a tracked git pre-commit hook, so every commit gets its own version with zero manual steps....
Talking about a beast, the performance of my text highlighting since we went full muse. We drown in resource BTW. I'm swimming in tokens like Scrooge McDuck.