how much did you pay for APIs during the 2 months?

does it use the claude code api or the claude code cli? You know, the claude code api is more expensive.

I also hope it can have a webapp version, rather than electron. because most of our work are on a remote server.

It actually uses the Claude code SDK so it plugs into whatever you already have.

It can use API/CLI or even if you have a private hosted instance.

We're actually working on a remote web app version but its a little trickier to wire up.

These are great questions - thank you!

If it uses the SDK then it's token burn? Or can it "legally" use your Claude.ai MAX account, your subscription account?

Great question - we're going to look into this in-depth.

Not allowed? This is easy to find in the Agent SDK docs

"Unless previously approved, Anthropic does not allow third party developers to offer claude.ai login or rate limits for their products, including agents built on the Claude Agent SDK. Please use the API key authentication methods described in this document instead."

https://platform.claude.com/docs/en/agent-sdk/overview

We saw this, but thought it was for Agents calling the API directly. Outworked is just a wrapper around your CLI using the existing agents and sub-agents in your Claude Code installation.

It's a great point though and we'll need to read into this more in-depth. Appreciate you raising this.

Right, pretty sure people got their accounts shut down for doing this kinda thing in the early days of OpenClaw before it was renamed and when it supported login. I guess you can take your chances but API is probably the only safe way to use this.

is this accurate? it doesn't seem to count the state tax?

I'm curious how a gpu language's syntax design can be different from CUDA kernel?

Because I think there is no way to avoid concepts like thread_id.

I'm curious how GPU programming can be made (a lot) simpler than CUDA.

Most GPU work boils down to a few patterns — map, reduce, scan. Each one has a known way to assign threads.

So instead of writing a kernel with thread_id:

  let c = gpu_add(a, b)
  let total = gpu_sum(c)

The thread indexing is still there — just handled by the runtime, like how Python hides pointer math.

how to prevent others from building a copycat using ai?

The discussion here is going sideways, and I blame the underwhelming blog post.

Having money is NOT an economic moat-- i.e., a durable, structural competitive advantage.

He overlooks broader, true definition of moat attributes like labor supply, infrastructure, PP&E, brand, network, natural monopolies, switching costs, regulation. These don't go away with commoditized CRUD apps.

And quoting someone with decades of experience implying that things are hard now and innovation didn't turn over industries in the last 25+ years is a joke.

The more I think about it the more brainrot the article really is. As if all problems are solved and pumping out soulless shovelware companies is worth anything. It really is "just one more app" all over again.

Knowledge ? For b2c it might be more difficult, but in b2b, understanding your customer and their specifics issue and developing something made for them is one of the big challenge. Being able to spit out code for free is useless if you don't know what and who you are making the code for.

The same way you prevented this previously. Copying successful products is nothing new, AI just makes it easier.

Marketing, lawyers, good customer support, creating relationships with customers.

You work on niches that have very specific requirements that you can only derive from having a good relationship with customers and so you attend to those needs faster than competitors who are out of the loop.

> you attend to those needs faster than competitors

I wonder if this type of hustling can be called moat building?

Let's just say, building software alone is not enough.

You cant. You can only focus on building your own product and making it durable and just much better

by keeping the how part a secret

You don't.

can you workaround the browser tab memory limit (4gb or 16gb) to process long clips?

Narrator: They can't

Edit: But that's fine, it doesn't have to. Both web NLE and desktop NLE can coexist. I'd find this useful for quick video editing on the go for memes and stuff, tbh.

tried the link:

Restricted Access Developer tools are disabled on this platform to maintain system integrity.

If you believe this is an error, please contact your administrator or return home.

You're right to be skeptical. The 'Restricted Access' was an automated runtime guard meant for our enterprise internal builds that leaked into the public demo. I’ve just disabled it. You should be able to inspect the network and WASM instantiation now.

similar to https://www.nimo.space/

Thanks for bringing to my attention. I haven't tried but I can spot some differences in philosophy:

- Musel Cloud doesn't embed Google Docs or Sheets, it has its own text engine; the word processor is built-in. There's no reliance on Google for docx, xlsx editing. Musel's built-in, natively on canvas. File formats such as docx, xlsx are compatible with any other software.

- Musel renders everything using a canvas (and WebGL), it doesn't use HTML at all. It is truly a whiteboard through and through. Everybody else has to rely on HTML for their rich text layout. Musel Cloud doesn't rely on third-party apps or popup a separate modal. Because everything is entirely natively rendered using the canvas, layers work really well, zooming works really fast, and performance is great. Items on Musel are more like Photoshop objects; they're all raster.

- has a built-in drive that syncs (but only supports Windows right now). No reliance on Google Drive or Dropbox.

Musel delivers on the promise of seamlessness. Works on files you already have on your devices.

I prefer IDEs, they have better support for me to drag and drop my screenshots to the ai agent, my projects are ui heavy.

I'm the same

I built something like ssh, but using a p2p protocol.

I can remotely control my home server, do port forwarding, copy files etc.