I don't agree. There's a limit on how much you can influence things, and as a dev (or product owner or whatever) you need to pick your battles.

How a decision is made, what gets released etc. are all opaque processes and blaming the devs from the outside is not productive.

That said, I'm all for it for devs becoming more responsible for the end result, evaluated within the org boundaries. Especially these days when people tend to cop out with "oh cloude did it / didn't catch it" responses (before, it was "oh I had to come up with something quick and couldn't check everything").

I use only providers because I don't have the hardware, but if, and that's a big if, the only moat left for humans is our ability to steer and make bigger picture decisions with our personality, I think it makes sense not to want to send all that conversation data to a 3rd party.

I use Opus for planning, through API on medium thinking, it doesn't cost anything significant and the rest can be happily done with deepseek or glm flash. Using sol 6.1 for review loops at the end. All running on pi.

A very complex task tends to cost like 10$ max. It's comparable to a subscription with enormous flexibility.

It's a steal.

I'm using DeepSeek a lot, and I'm curious how you get a task to go to ten dollars with it? Or are you mixing Opus and Sol tokens to the same task? I have very hard time to go over 5 dollars with DeepSeek and that is already a weeklong project where I share the same context.

Of course I do certain things like use omp and do a compact picture snapshot after 256k tokens automatically. My tasks rarely go over a dollar...

I've ported a few games, wrote a cosmic desktop app, iterated in my own agent software, use the model in our home speaker system, use the model in my phone assistant. My monthly bill to coralbricks is about 14 dollars.

That's the total cost of a task across all models. DeepSeek going above 2-3 dollars is actually rare because of the very low cache costs.

I also used to plan with more expensive models like you did, but now I just do planning with DeepSeek too and don't see much difference. I use the omp plan mode and iterate on the text until I'm happy with it. Then clear context, give the plan to DeepSeek again for implementation.

If it's mission-critical code, do a security review with GLM 5.3.

Otherwise go manually through the code and ask for changes if needed.

Works really well for me.

I'm curious why you use Opus for planning and Sol for reviewing, instead of the same for both? Do you find they are better for those tasks?

Thank you deno!

You are added to the list of projects I thought would really become essential to my work but slowly faded away. Another one, for example is rethinkdb.

At least Deno actually caused some movement in the node world, just like io.js.

Very cool idea, I'm going to test this!

One nit: Forgive me if I'm missing something but why do you have pro components? Won't people vibe-code clones anyway? I'd rather have a donate button these days.

Hey! Author of Proa here, thanks for checking it out! The idea is to just accelerate the time it takes to create components, similar to Shadcn or other component libraries. You can create a new component and modify them however you like. It’s nice because they have both the WebRender and MdRender trait implemented, and saves you a bit of time. Also for slightly more complex components like Calendar, we’ve crafted the component with all the JS interactivity required to do things like switch months, select start and end date, etc

You answered "Why do you have a component library?" but I think the original question was more along the lines of "Why bother monetizing components that can be copyright-washed in minutes with LLMs instead of having a Donations funnel?"

In any corporate structure it's 10000x easier to get approved to just buy something from a vendor than it is to set up donations for an open-source project/foundation.

PLus you're missing the licensing aspect, part of this is AGPL if you use the free version, lots of places are happy to throw down money to get a commercial license rather than hem and haw over AGPL.

Aren't Pro components the same as a donation, but then giving you something extra as well.

Kind of, not really, in this case it's a subscription that renews monthly/yearly so you have to remember to cancel it when your reach the amount you wanted to donate, maybe you didn't want to donate as much as 1 month costs so you're left not sending them any money in the end, right now you can't even subscribe cause their console 404's. You got to make it as easy as possible for people to part with their money, component library subscription is not it.

I vibe-coded a paste manager using wails (https://github.com/egeozcan/mahpastes), and I'm currently on linux (bazzite) and I have 0 problems with the rendering.

One of the things I hit is that there is no way to trigger an "open folder" dialog. It gets mapped to "open file" instead. The correct hooks exist in both WebKit and in GTK, but not wired up in WebKitGTK. I think the fix would be quite minor but I looked into what is involved in raising a patch and decided to leave it. There are a lot of minor paper cuts like this.

I did hit a couple of CSS or rendering quirks, but the rendering perf also seemed noticeably poor, and that's harder to live with.

Try it on an older LTS distro release that is still under distro support or even better, one that's under extended paid only support. Anything pre 2022 is probably completely broken and anything pre 2024 might have strange bugs. There was a great one where building a Tauri or Wails app on Ubuntu 22.04 for better compatibility but running it on a newer distro had a font rendering bug because of the older Webkit2Gtk. Trying to display emojis crashed the apps.

Recently vibe coded this https://www.rominabox.app/ and also loved Tauri for the "builder" part of it! What a pleasure to work with.

Granted, it's very simple, but I felt it was a breeze.

If a teacher had to hold my kid as a last resort, instead of being mad at the teacher, I'd be motivated to do better on teaching my kid how to behave, and apologize from the teacher.

You also need to restrain kids sometimes (could be often, if for example they are in the spectrum), because they have trouble regulating their emotions and they can hit others, and even themselves in crisis situations. However, you should explain them later, why that was necessary.

The teachers not being allowed to do that is mind-boggling to me. What do they do now, bring their parents with a helicopter?

Sure, it's always easy to imagine the "positive scenario" (the kid really did something wrong and got lightly restrained).

Now when you authorize a restrictive / possibly violent behavior on people having authority, and you will default to "believe them" (because they're the literal "adult in the room") you also open it to abuse

See what happened with police in the US...

Do you have a link to the bug?

BTW. The "won't load new pages" may be something else. If an agent launches its own Firefox in a new profile, it will sometimes trigger the auto-update which will overwrite the binary in the profile you were using. It won't crash but you'll get a warning to restart.

Why would you give an agent rights to overwrite system wide binaries?

It shouldn't even have access to the same binaries you are using but only those within the jail it is running in.

It doesn't need rights to overwrite system binaries, it just needs enough permissions to tickle the updater. And the updater has the rights, for obvious reasons.

I don't know, firefox typically do not comes with an updater on linux as it is updated by the package manager.

Some valid points but I'm not sure if I agree with the brush engine point. I don't think it's the most important feature of Photoshop to many, so not making that one right first is not that surprising.

Making a good brush engine is indeed a very hard problem, especially when you have undo history, high res canvases, layers, zoom etc. I just don't see it of such importance

Which is why you aren't the audience of people that pay for Photoshop.

My team are and the full Adobe plans and for our use cases it's also not important. We do have a few workflows that dive deeper into niche tools only Photoshop has though, but I can see competitors making streamlined apps for specific use cases that can compete. It's actually kind of tempting to see if I can get something done over a couple nweekends and see if it'd be usable

The PhotoCraft repo already appears to have over 200 contributors.

I'd consider adding/improving the features directly there, unless a more streamlined alternative is the goal.

I wonder if the project only takes contributions that aim for parity with the Adobe software or also new features and improvements.

That's very superficial. I used to pay for Photoshop and I was very happy until they started with the subscription business.