oneneptune

"just ticking up a version number" drastically understates the headaches of Material UI over the years.

Not a boring application, a very large application -- each major version update was a tedious process because they completely upended many APIs.

We've migrated to shadcn, and upgrades are now easy; we can upgrade a single component -- add the new ones free.

With MaterialUI we had to update EVERYTHING to their new APIs to be able to take advantage of the new features anywhere.

With shadcn we can be selective.

All the independent cleaner's I could find wanted $400 a week to clean. I've tried 2 different cleaners and fired them both after 4 weeks. They only spend about ~3 hours here a week, so I'm paying over $130/hr... and they didn't even do that great of a job. I guess I should try a third? This robot is tempting though.

... they pay low income people $2-5, literally, to be the face of their scalping. It's like gig work for those people. The worker uses their own name, address, and credit card even to make the purchase. The scalper reimburses them + $5.

... and building a business off of skills you don't have based on a strategy already exists! You use capital to pay humans that do have the skills.

Or capital a comparable sum to pay an AI to approximate the skills of humans I guess is the proposed future?

I do not really like applying the "if we did it, they will too when they can!" logic to other government's.

China has flaws, plenty of them, but there's no real evidence to believe their motivations or mechanisms of pursuing motivations are that similar to that of the United States.

Don't worry, China and other countries won't be so dumb with their models.

Wow what an awesome art piece by Ada Zejun Shen that they commissioned(?) for this article!

I'd love to do all 3, and actively push my city council to prioritize rezoning for mixed use + medium density.. and eliminating parking minimums. I previously lived in NYC and it was true like the author said; I could have boundless weekend trips to a variety of places with Amtrak + Bus service.

This is great unless you live in an area of almost absolute geographic and social homogeny in a 100 mile / 160 km radius. "Yo friends, want to drive an hour and see if the fast food in a strip mall is the same as our fast food in a strip mall" just doesn't quite land or "Want to drive 45 minutes and walk in a park that was built in the early aughts and lacks proper shade and had all it's benches removed just like ours?"

The uncanny valley of text. It looks and sounds like a human, but lacks the "soul" / humanity that our intuition somehow perceives.

It's really strange... I see some text with obvious tropes and sometimes I read something and there's no obvious AI trope... but it's just not human?

I once saw the bad side of one of these draconian state laws many years ago. People rarely have the misfortune of hitting these laws in some flyover states... and I remember the local judge being really shocked by the mandated penalties for such a simple offense.

idk I just like running 6/8 terminal panes and organizing my workflows / projects in an exact space. I even tweaked my theme. and seeing them all on my side portrait monitor.

This is presuming the next president isn't supportive of these policies and a continuation of the policy.

The biggest problem is the author didn't even bother to verify it. I've seen big multi-model ralph wiggum or whatevers do a conversion. Run for 6 hours. Upon manual inspection to understand how it handled some tricky calculations / logic I find stubs and hard coded truthy returns. So even if you ran a smoke test suite against it -- you'd think it successful...

The value in paying someone is if you have enterprise requirements for physical data security. Then after that if you go the Hetzner route, you have to micromanage your underlying OS, Redis, DB, etc... and it's just more work.. and if you're in enterprise business it reduces friction a lot to just pay someone a trivial amount like $10,000 a month.

anecdotally; this is why I now use fb marketplace exclusively to sell stuff. cash only. meet in person; feel free to test / inspect stuff. All sales are final and are as-is.

... and speed. They publish billions of ticker items in near real time simultaneously to all their users... the tech is pretty crazy!

I think given the scale of projects worked on / being estimated (8 to 9 figure projects) -- it's unlikely we'd be willing to test on our real CAD files. Any chance for enterprise hosted solutions that aren't on the cloud?

Please forgive me as I'm both naive and ignorant;

What is the implications / meaning of that? If a software comes from a trusted source and there's adequate mechanisms to make sure file contents aren't stolen / accessed -- why be concerned with who else has the same software?

Also, in the construction industry you get an updated drawing file a day before the bidding closes... good luck getting the GC to send more detailed files (that they themselves got elsewhere) in that time. You're better off sending it to your estimation department in India and letting them work through the night to put together the new estimations.

Is this a service / product you plan to offer outwardly? I'd be interested in learning more. Use case: estimation.

This is very cool.

Anecdotal solely to me, very unnerving and even with formatting marks enabled makes me feel uneasy seeing a space without a space formatting mark.

Courier Prime won for me, I've always been a courier fan I guess because I wrote all my books in the 90s with it..

A personal anecdote:

I had several roommates, and we each were responsible for a utility. I was responsible for internet, and Cox was our provider.

I received multiple e-mails from Cox about copyright infringement. I can't recall them, but I remember it being serious enough for me to tell people to stop.

Thinking back, I feel like Cox's position is right and fair; let users know they're being observed by copyright holders, and inform the user that they could be compelled to provide their identity to complainants.

But ultimately, the responsibility to "stop" the supposed infringement is on the holder, not Cox.

One very minor note; Anthropic and others, like most "enterprise" solution, also sell SSO + SCIM + audit logs. Their business plans have lower tokens and higher prices to cover the enterprise features, which should be essentially free to provide in 2026.

I think people will be quick to engage with the "ai is risky" angle, but the thing that jumps out to me is that you were working against a production state in the first place.

The agent made a mistake that plenty of humans have made. A separate staging environment on real infrastructure goes a long way. Test and document your command sequence / rollout plan there before running it against production. Especially for any project with meaningful data or users.

a redis geospatial index + redis distribution locks you can build a performant cache layer that is consumed by a ton of people and stay well under that rate limit.... the weather data only updates every 5 minutes too, so you can use that for your cache ttl.

Nice, you fed bullet points into the LLM to make a book. I can stuff your book into an LLM to make bullet points. It's like a wonky non-deterministic hash for both of us to save time and give the author some feel good dopamine at the expense of tokens + authenticity!

I'm a Claude Code user primarily. The best UI based orchestrator I've used is Zenflow by Zencoder.ai -- I am in no way affiliated with them, but their UI / tool can connect to any model or service you have. They offer their own model but I've not used it.

What I like is that the sessions are highly configurable from their plan.md which translates a md document into a process. So you can tweak and add steps. This is similar to some of the other workflow tools I've seen around hooks and such -- but presented in a way that is easy for me to use. I also like that it can update the plan.md as it goes to dynamically add steps and even add "hooks" as needed based on the problem.

Thanks in advance, I can't wait to see your prompts and how you architected this...