nkko

If they pull it off, ChatGPT becomes the standard UI for 50%+ of web usage. If policy fails or tech execution lags, they’re just another chatbot.

Yes, and one test failed as it missed pydantic dependency

this is very cool considering it is a fine-tuned 1B model

I didn't realize that it wasn't attributed; it was written by @AlpinDale.

Morph.so is an ai sandbox thing that has branching from state, prob doing a talk in SF on this at AI Engineer conf

As far as I understand Firecracker, you can't do a GPU passthrough

Check this out: Spinning 100 Agents in Daytona https://youtu.be/OFFFyfgO2ik we should also soon release open source speed benchmark

The single biggest growth hack is getting that product market fit.

Thanks for the share. I would love to hear the community's opinions on the structure of the mapping (layers and sublayers), as it is always very hard to reduce dimensionality while keeping enough meaningful structure.

Another AI social media bot. For every one we see on GitHub, ten more lurk in the shadows, cheerfully automating away what little authentic human interaction remains online.

Soon we'll have AI agents optimizing engagement with other AI agents, while real humans wander off to have actual conversations somewhere else - probably in a cave, using sticks to draw memes in the dirt.

An alternative approach we're working on is Daytona (https://github.com/daytonaio/daytona) - it orchestrates dev environments on your infrastructure or even on your local machine. Oh, and it is Apache 2.0. The interesting bit is how it handles any IDE integration and various providers for running your dev envs on.

Disclosure: I work on this project

We are having some trouble atm due to the incoming traffic from the PH launch, but we hope to sort it out soon. Btw thanks for the share. Open to answering any questions.

I skimmed it without even realizing the date of the article. It's remarkable how even content on dated tech like this still feels fresh due to the same underlying problems.

Could we work toward standardization at some point? Obviously, there will always be a newer model. I just hate that all the embedding work I did was with now depreciated openai model. At least single providers should see interest in ensuring that for their own model releases. Some trick like matryoshka embedding could secure that embedding from newer models nest or work within the space of older model preserving some form of comparability or alignment

vector embeddings of the songs plus coop network

That would be a fork without wifi manager. ;)

trivializing their purpose and the suffering they represent is selfish, the act of skateboarding on these monuments is deeply troubling

Authorization is such a critical yet underappreciated part of system design. It's great to see open source projects tackling this space. I'll definitely be checking this out for my next project. How does this handle multi-region deployments?

How about dedicated external system for this, running on rPi or proxmox instance. On the local macgine app should only record screenshots on major scene change and save them on remote fs where it would go tgrough ocr, indexing, embeddings, whatnot.

I stoped reading at “in the realm”

Ooo, printing this for the kids room.

I'm here for the phenomenon of the GitHub stars on this repo. Which was quite a weird event. Even with just the readme in the repo they got a crazy number of stars and were on GitHub trending for a day or two. What's with that? Are there just so many front-end developers, is it possible that DND is such a huge unsolved pain, is it just cause its Atlassian, did they sent a mass e-mail? I am honestly confused? What's the value of stars these days for an individual developer consuming the code?

I just host LibreChat on US server.

This is the reason why I always get back to HN. 100% value delivered.

The logic is flawed but the reason is reasonable.

Application of AI such as this gets me excited. Untangling complex logic and turbocharging boring tasks. Authorization systems can get very hairy.

Having an AI assistant that can spit out a well-designed model and policy to implement it could be a huge time saver.

I'll have to take it for a spin and see how it handles some real-world use cases, but I'm excited to see tools like this pushing the boundaries of what's possible with AI-assisted development, especially in gnarly domains like auth. Best of luck with the project!

Hey Melanie, thanks! After working on this project behind closed doors and, now going public, we've become hooked on the positive feedback.

Thanks again for sharing your thoughts and experiences with Product Hunt launches. You make some good arguments there, and I haven't thought of it that way. But long-term visibility could be a win for sure.

Focusing on building a great product, gathering meaningful feedback, and leveraging multiple channels for promotion and growth is key.

Ah, I see. That clarifies things. Until today, I was a user. But this change in roles brought a perspective change.

Honestly, we are lucky to have one of the founders bring his huge network of contacts from all over the world, as he has organized large events in this space and was a CDXO for a worldwide company.

Everything already exists, which is no excuse for doing it again.

We’ve ramped up our outreach because experience shows it’s critical for building the contributor base and fast-tracking project evolution—our flurry of repo activity is a testament to this strategy. On the email front, a technical snag during our provider’s migration led to the mix-up where some subscribers were left enrolled. We’re on it. Meanwhile, we believe Daytona’s as a product will make up for the noise. Check our Friday release v0.7.0 for the proof in the pudding.