krts-

Is there a transparent way to see credit used/remaining/topped up, and do you have any tips for how you can prompt the agent that might offer more effective use of credits?

The LLM chat taps out but I can't find a remaining balance on the fly.io dashboard to gauge how I'm using it. I _can_ see a total value of purchased top ups, but I'm not clear how much credit was included in the subscription.

It's very addictive (because it is awesome!) but I've topped up a couple of times now on a small project. The amount of work I can get out the agent per top-up does seem to be diminishing quite quickly, presumably as the context size increases.

Hero.

This is incredible. It does seem quite expensive compared to Zed or Claude Code now it's on Pro. But neat enough I've burned through the $20 subscription credit despite being a bit of an AI sceptic. This seems to have a much better handle on UI design (unless I'm missing something with the other agents), but as a solo dev I'm becoming quite convinced. It's also got me to try out fly again.

I couldn't get Tidewave working but I must try again to see if Tidewave with Claude Code would offer this level of awesome.

ps. @fly - please let me buy more credit, I just get an error!

A great project with awesome implications. Well deserved, and the fly.io team are very pragmatic.

This will be even more brilliant than it already is when fly.io can get some slick sidecar/multi-process stuff.

I ended up back with Postgres after my misconfigs left me a bit burned with S3 costs and data stuff. But I think a master VM backed by persistent storage on fly with read replicas as required is maybe the next step: I love the simplicity of SQLite.

I like modernc.org/sqlite[1] - a port of the C SQLite3 library to Go. No need for CGo, so straightforward to build and cross-compile.

[1] https://pkg.go.dev/modernc.org/sqlite

The headline is misleading. Mechanical CPR devices have been utilised for years in the UK. The major LUCAS clinical trial [1] was run in the UK.

The ‘first’ relates to the version used: LUCAS 3. This has no performance difference but adds ‘smart’ features via Bluetooth. This is also misleading, as the LUCAS 3 has been available in other services.

[1] https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

There's faasd [https://github.com/openfaas/faasd] which is OpenFaaS without the Kubernetes overhead.

It feels like SwiftUI, and other native tools, are embracing the advantages of some of these other cross platform tools: declarative, rapidly-iterable, abstracted.

I feel SwiftUI markedly lowers the 'barrier to entry'.

Crossplatform tools may serve your purpose better, but I think it's unfair to say that, at least for SwiftUI, native development may not be as 'joyful'. A list view in SwiftUi is List {}.

This sounds unreasonable, but it feels like a case where you don’t win either way.

Updating and changing rules reflects improving and novel knowledge. This is preferable to blindly following methods where disproved.

Rapid changes likely damage trust, increase confusion, and risk poor compliance.

A balance of the two is pragmatic, but it still seems to be new ground.

That's not quite fair.

Parler is listed at 2.8 million total users back in July.[0] Assuming that's doubled and all lean towards Trump, that still represents a fair minority of voters.

The website remains there for those users that want to use the platform, but curated app stores that aim to promote certain values (and are fairly clear about those values) don't seem to be the right place for these platforms that devolve into highly controversial places.

[0]https://en.wikipedia.org/wiki/Parler#Usage

A need for moderation was surely not a surprise.

Using a lack of moderation as your selling point and then being surprised by such a deadline when you intentionally host content that doesn't tally with the platform you're on seems short-sighted.

The intent is likely to not moderate - it'd kill the platform. It remains accessible by web. Does it need a native app?