we don't need machinery or a mechanism to compare it, natural selection works just fine for 99% of all species on earth.

All those things are beyond the demo itself. Vibe-coded demos are just demos. There are stability, security and everything enterprise that still needs to be added to a demo to actually make it functional as a paid offering.

The hard problems still remain.

My point is that if I were an investor in the LLM-era I'd be shifting my attention to the answers to those questions more than I would to tech demo.

I remember how Skype, an awesome piece of software transformed into Lync, which worked fairly well, slowly transformed into whatever MS wanted to call it year after year, slower and more buggy than the year before.

Lync started out as "Office Communicator 2007" before being renamed Lync. Then Microsoft purchased Skype and rebranded Lync as "Skype for Business" even though it was still a completely unrelated product, with just some basic interoperability slapped on. Skype-the-consumer-app lived on separately as its original product in parallel.

Just another example of how Microsoft is utterly incompetent at branding - always have been and always will be.

This is one of the most frustrating thing about working with NextJS. There seems to be no way to improve the speed of building the app.

I've used the other major meta frameworks (remix and tanstack). I don't think there is a way to improve the speed of building the app in those ecosystems. Happy to be proven wrong.

Could be the bundler re-resolving the whole dependency graph on every build, even when nothing changed.

I am unaware lm studio is being used for fine tuning. I believe it only does inference.

Happy to see unsloth making it even easier for people like me to get going with fine tuning. Not that I am unable to I'm just lazy.

Fine tuning with a UI is definitely targeted towards hobbyists. Sadly I'll have to wait for AMD ROCm support.

Thanks! We do have normal AMD support for Unsloth but yes the UI doesn't support it just yet! Will keep you posted!

What does "normal AMD support" mean here? I was completely unable to get it working on my Ryzen AI 9700 XT. I had to munge the versions in the requirements to get libraries compatible with recent enough ROCm, and it didn't go well at all. My last attempt was a couple weeks before studio was announced.

Is that not already possible with Markdown spec files and planning mode?

I guess? At least there you can review the plan, but is this planning mode any better at making architectural decisions than when you prompt an LLM and let it make the changes directly? (it might be, just not sure.)

You are right.

Although I have been a skeptic of MCPs, it has been an immense help with agents. I do not have an alternative at the moment.

Elaborate?

It seems to be a way to circumvent censorship.

They probably want to tone down their marketing claims then, since it doesn't let you browse the web without internet access. It lets you browse webpages that would be blocked in your area

I thought this was going to be some kind of wifi or bluetooth mesh

Me too, or some kind of offline caching system that automatically downloads your commonly visited websites or something

I had the exact same thought and it wasn't until navigating to their FAQ did it become apparent what they were actually selling.

"Browse the web without internet access" is either purposely misleading or written by someone who doesn't understand the tech they're marketing.

in Iran there's a funny situation where you're connected but not to the internet. if that makes sense.

Are you still able to access Iranian servers, or is all traffic blocked? I'd class "The servers in Iran" as still a part of the internet

It does thanks. Hope all is well if you're over there.

I agree. "Cut off" and "without internet access" != censored. They should have worded it better. The title is the worst.

"Ceno, browse the censored web if BitTorrent isn't censored."

I absolutely love it.

Am so much more excited about tiny models gaining real intelligence. Just today I have been running Qwen3.5 0.8B model on images and am pleasantly surprised by how good it is compared to even 4B and 8B models from a few months ago.