The goal here is to make it easy to bundle applications completely in a browser.
The primary use case is to support AI generated applications at run time.
For example: prompt an AI model to generate a full application including its package.json and all of the components as "files" and then using a virtual file system and import maps bundle it and run it.
This is a key piece of the puzzle for how a vibe coding platform would be built. Companies like Bolt.new have proprietary technology that runs node in a browser to handle this use case.
Which is very cool but not accessible to everyone.
I am not sure how Loveable does this... they may have a similar library to this one. I'd love to know if someone wants to share! :-)
This library was itself vibe coded using Claude code. This is my third attempt at this... the BIG key was really honing in on the exact requirements (LOL... that has always been the key right!?!).
I'd love some feedback if you have any. I am sure there may be some other option out there for this but I wasn't able to find any easy wrappers of esbuild-wasm myself that fully solved my use case.
Now I can go vibe code the full vibe coding platform!!
I am a long-time software developer too... but I am strongly encouraging people to embrace the future now and get creative in finding ways to adapt.
There will always be demand for smart and creative people. BUT ... if you don't look up and look around you right now at some point you will be irrelevant.
Also see Ray Dalio's "Principles" book on embracing reality. A critical principle in the modern age of AI.
I actually thought this was pretty boring. @Cursor team.... hit me up when you do your year in review for 2026 next year and I'll help you brainstorm something much more creative.
I entered the topic of "going to war" and it did not return anything, but throughout history people have used the Bible as justification for going to war. And yet killing people is generally bad. Lots of nuances here but it seems like the system should have come up with something.
Nice! I am currently writing a new version of my no-code platform, WeBase [1], to use AI to generate and edit applications.
Currently just using foundation models from OpenAI and Gemini but will be very interested to try this out.
My current approach is to just completely overwrite files with new updated version but I am guessing using something like Relace will make the whole process more efficient... is that correct?
I'll watch your video later but I would love to learn more about common use cases. It could even be fun to write a blog post for your blog comparing my "brut force" approach to something more intelligent using Relace.
This looks really nice. There is a pretty bad UX bug though... when I click export I can't see that anything happens. But if I scroll down I see there is a popup to export the chart. The popup really needs to be visible in the view pane so a user knows something is happening.
Yep. Just added Deepseek yesterday. The R1 model never returns output via the API so it is currently disabled, but you can access the V3 model even with the free version.
I built NimbleAI as a way to compare how different AI models work.
Currently it supports all of the usual suspects: o1, Gpt-4o, Claude 3.5, Llama 3.2 and Gemini 1.5. I also added support for Deepseek V3 recently (their R1 model seems down right now so it is not enabled).
The ultimate goal is to build agents that can leverage the best AI model for the job to complete more complex tasks, but to start I built a foundation for experimenting with multiple models.
You can also make chats or chat groups (comparisons) public. Here is an interesting one comparing Gpt-4o with Claude Sonnet 3.5:
This is nice! It looks like the pipeline is filling up! I signed up to launch NimbleAI [1] on your site.
One small issue with the launch date calendar. When I clicked a date in the calendar the popup with the actual launch date was always one day later. I wanted to launch on Tuesday, Jan 28, but when I picked that date it scheduled the launch for Wed, Jan 29.
BTW... what was the product you launched on PH that got buried??
Also... how funny would it be to launch ItsLaunched on PH and have it bubble to the top of PH! You should coordinate a launch on PH with all the users signed up for your new tool to get lots of votes. THAT would be awesome!
I had fun writing this. I really didn't like TailwindCSS when I first looked at it, but came to really like it after I worked with it enough to "get it".
I also love the basic ideas of this article but think the discussion on "micro-management" should be revised.
What about the idea of "Coaching Done Right" with some discussion of how to mold and encourage direct reports toward the path that leads to optimal growth for them?
The term micro-management is an inherently negative term. It connotes the idea of hovering over a person and trying to correct every detail of what they are doing "wrong". It takes away their agency and ability to learn from their mistakes.
Love the core idea of the article though! Thanks for sharing.
I would start by asking "What is my purpose in seeking to make a switch?"
Knowing your 'why' will help you know for sure if you are able to make the commitment required to make a switch. In general it will take a non-trivial amount of effort to make the change.
If you are confident that you genuinely want to move to the new field then it is all about learning and growing the skills required.
Get 1% better every day in that field and within a reasonable amount of time you'll have the proficiency to at least get entry level position.
It is completely free to try out and use. The deal gives you the integration with Netlify to publish to custom domains, the use of the AI Assistant, and the ability to download the source code.
Actually we haven't put the download behind the customer features yet so you can do that today, but at some point that will be a pro-feature. Of course developers can just inspect the source and copy anyway so it kind of is always "downloadable". :-)
Yeah... there is always more to the story than what we see on the surface... and it was an aggressive comment.
Honestly I knew the missing links in the footer was lame but just wanted to get some feedback and was tired of delaying "talking to potential users"!!
Now I will have renewed energy to fix these things!
And frankly I am glad I did... it made it to the front page of Hacker News for over two hours so at least people are interested even if it needs more polish!
The primary use case is to support AI generated applications at run time.
For example: prompt an AI model to generate a full application including its package.json and all of the components as "files" and then using a virtual file system and import maps bundle it and run it.
This is a key piece of the puzzle for how a vibe coding platform would be built. Companies like Bolt.new have proprietary technology that runs node in a browser to handle this use case.
Which is very cool but not accessible to everyone.
I am not sure how Loveable does this... they may have a similar library to this one. I'd love to know if someone wants to share! :-)
This library was itself vibe coded using Claude code. This is my third attempt at this... the BIG key was really honing in on the exact requirements (LOL... that has always been the key right!?!).
I'd love some feedback if you have any. I am sure there may be some other option out there for this but I wasn't able to find any easy wrappers of esbuild-wasm myself that fully solved my use case.
Now I can go vibe code the full vibe coding platform!!