@PovilasKorop I want to think BIG but damn, there are so many little details out there that would be probably solved in a minimal amount of time if they were treated as what they are. A major part of a great big product.
@PovilasKorop I dream with the day Alexa's Spotify skill allows the user to have one or multiple loaded profiles independently on each Alexa device. Apparently, AI is not enough for maintainers to take a look and fix that so far. Does this count as big since we are talking about 2 big cmpns?
@PovilasKorop I have found myself crafting full design and identity systems more and more every time. Just leaving very final details (undetectable at first glance) to the implementer.
It's kinda funny trying to get this designs almost tech-agnostic. Just letting them know foundamental tech.
@PovilasKorop I know it has almost nothing to do with coding but it is technically related to AI and agentic tools. And solo developers sometimes want a cool base to take inspiration from during its implementations.
@PovilasKorop Is this post any kind of marketing for your Native PHP app Povilas?
Haha. No but seriously, Idk how many times I arrive home and suddenly remember why I leave in first place. Embarrassing at my early 20s.
Google Keep and WA's autochat are useful sometimes.
Hope this finds its way to the right people.
Any thoughts on adding infrastructure settings mounted within Mexico for @laravel Cloud? Working for government stops you from choice the perfect cloud due to legal restrictions (data should not live out of our borders).
Hey @aarondfrancis hope you see this. Idk if you are aware but when using Solo on languages that handle accents and you type as common as you do in that language the accentuated glyphs dupes themselves. Also cursor movement with cmd/opt+arrows on CC cli is overridden by Solo keys
@thsottiaux@paper I don't know if it has something to do with the sub tier (I was on Plus, ending today). But I wouldn't get the Pro plan just to figure out if it works better. I literally have a "Codex fails" folder in Paper where I put all the iterations of the same product I did with Codex.
@thsottiaux It's pretty bad for design tasks. I mean when you use the term Agentic you mean it should work agnostically through any tool applying the same principles but I haven't been able to make it work as good as Claude with @paper and the design loop I use there.
Since I don't like this kind of workaround (as the proposed by the model) I simply configured manually as indicated in the @paper docs. Anyway, the research done by the agent resulted in my 5hrs limits dropping from 96% to 45%, what a day.
@paper and @thsottiaux any thoughts on this?
Paper MCP works well in any other tool tool except by codex ones. I already did what any other human would do: I uninstalled the plugin, then uninstalled the marketplace from codex and reinstall everything again. Thanks in advance.
@PovilasKorop Suspicious. I remember to have seen this post yesterday but I'm seeing it now and it says it was posted just 2h ago. Good package btw, it reminds me of shadcn/ui but for Blade.
@thsottiaux I just applied this and made 3 one-line propmpts and my limits went from 100 to 96. Are you kidding me?. I have never used/abused of api usage as mentioned in several posts I have seen recently. The only way I have used gpt models is through this and OpenCode (auth not api).
@PovilasKorop I have been seeing the model looping itself over and over again for both mid-complex and simple takss. I hadn't catched that it was because of effort level lol. The bad thing is that I often need high effort to produce usable output on design related tasks, but I'll see