Giovanni

Web Developer Freelance

[Giovanni @Giovi1905]

@thsottiaux Thank you. Sol 6.1 is a real disaster when it comes to frontend design, especially with padding and margins. Are you doing anything to address this? I’m forced to keep using Claude for this.

[Giovanni @Giovi1905]

@designwkarthick 32GB, without a doubt.

[Giovanni @Giovi1905]

@Lbdev__ From what you’re saying, it’s clear that you’re using AI the wrong way. Nothing could be more inaccurate than comparing AI to a stupid intern.

[Giovanni @Giovi1905]

@thsottiaux It is very reliable. Only rarely have I encountered coding or interpretation errors during post-change testing.

[Giovanni @Giovi1905]

@ivanfioravanti Yeah, it doesn’t seem to be MiniMax H3 for Apple Silicon. What’s the best experience you’ve found for AI video generation on Apple Silicon chips?

[Giovanni @Giovi1905]

@thsottiaux Please no 5 hour limits!!!

[Giovanni @Giovi1905]

@precisox

[Giovanni @Giovi1905]

@antirez Same experience with Q4.

[Qwen @Alibaba_Qwen]

Qwen3.8 is launching and going open-weight soon!🌐

With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5.

You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba’s Token Plan, Qoder, and QoderWork. Be among the very first to try it out.

Can't wait to hear what you build. Stay tuned! 🚀 

Token Plan
international：https://t.co/YRvcGdB9Bv
China：https://t.co/PKMUNwUuRp

[Giovanni @Giovi1905]

@thsottiaux Without the 5-hour limits, it’s perfectly fine as it is. Each of us can manage our own weekly limit.

[Giovanni @Giovi1905]

@antirez @matteocollina The real fear is admitting that writing code manually is no longer the point.

If tests cover the logic and everything works, why should I review every single line by hand?

[Giovanni @Giovi1905]

@PovilasKorop I have to say, with DwarfStar by @antirez I can run DS4 Flash locally, and paired with Claude Code + Laravel Boost I’m getting good results.

[Giovanni @Giovi1905]

@mobailabs @ivanfioravanti I’ve tried everything: DS4 Flash, Qwen 3.6, and more. Maybe my workflow just isn’t a good fit, but I still find local LLMs too unreliable for serious development. What local LLM do you use?

[Giovanni @Giovi1905]

@antirez I keep asking myself: for projects that need to go into production, what’s the point of using these models if you still have to run everything through a frontier LLM anyway, otherwise you end up with tons of bugs?Isn’t it just a waste of time?

[Giovanni @Giovi1905]

Honestly, I’m feeling pretty discouraged about using local LLMs.I have an M5 Max with 128GB RAM, I’ve tried almost everything, and the best model I’ve managed to run locally is DS4 Flash by @antirez.At the end of the day,

[Armin Ronacher ⇌ @mitsuhiko]

Europeans please wake up. https://t.co/vBIdm5ci9k

[Giovanni @Giovi1905]

@dscndt @Egomnia Non è possibile, penso che sia una trovata pubblicitaria!! Non posso credere che tutto sia realmente così assurdo!!!