prologic

I'm not really sure I agree entirely with the premise that plaintext files are going away or becoming less prevalent. I tend to do a lot with plain text files myself, but maybe you're right, maybe I'm kind of old :)

To be honest, this is less about Claude Code per se (or any other coding agent) and more about business and corporate culture. In my experience, if you work at a company whose culture is to deliver as fast as possible, then this is ultimately what you get when those same managers start to realize we can also ship 10x faster (no matter the costs). The "Delivery Pressure" is very real in _some_ corporations, and I _believe_ things like Claude/Codex just accelerate this even more, which is what we're seeing in some places. In a fast-place industry, can you blame 'em?

I assume you're talking about this NDIVIA Spark DGX[1] ?

It isn't quite the question/answer I was hoping for, but anyway. Thanks!

[1]: https://www.nvidia.com/en-au/products/workstations/dgx-spark...

Hey folks

This is something I had been working on for a while now (few years, even before things like Claude/Codex were considered "mainstream").

It's basically a Linx from Scratch (LFS), with a twist. Every part of the system (except the Linux kernel itself of course) is written entirely in Go. I've done similar things in the past with what GoNix is derived from, called uLinux, and some shell-script based LFS (but widely different then the book).

Anyway, GoNix is pretty neat (IMO), it runs https://gonix.dev itself! And it boots in roughly ~20ms in Vultr Cloud where I run the site from. I'm also using GoNix as the base OS for an appliance I'm building with OTA and A/B updates. So this isn't just a toy, it's quite serious really.

Target market/use-case is appliances and use-cases where you need a system to be quite small and quite fast to boot-up, e.g: Ephemeral CI/CD runners.

Sorry, but what's a "modset"? I _assume_ you're renting GPU(s) from a IaaS provider?

I'll say this again, why must every technology and every product we ever build devolve into an Ad-serving platform?! Farrrk me.

What's your cost model / pricing based on here? If you're going to provide $20/month for unlimited APi calls, I can see this getting abused very reasily.

Very nice updates Umbrel!

Good timing too, Just as xiaomi[1] makes their big announcement :D

[1]: https://news.ycombinator.com/item?id=49792730

I will stop using ChatGPT in 2027.

I think a lot of folks ended up in a world where you write specs and designs in Markdown documents and use that to drive the outcome (the code, the system, tests, etc). So you drive intend through those specifications and design docs.

Lately though, I've switched tact a bit and have resurrected some ~20yr old research called Behaviour Tree Engineering or Behaviour Trees, so I'm now writing specifications called BT specs.

e.g: https://bte.mills.io/s/v4Mjbyf0OBJWm5YRLwG5XiIH

Same. I don't get it.

Oh wow! :O I always remembered this being called "The Sabbath". I had to go look this up. But oh my, wow! We're not starting to talk about AI and Religion in the same breath?!

Because consistency is still considered a quality rubric ?

I think you have to rethink this a bit if I'm to be honest. All the login/signup pages says is:

> There for you, every step of life. > > A galaxy full of paths — we'll find yours. Explore, find your fit, and chart the life you want.

This tells me nothing.

I'm sorry but it took so long to load, and when it finally finished, it presented me with basically just a login/signup page. I just closed the site. Sorry. Not for me.

I agree, precisely. I'm kind of sick and tired of it to be honest. What if we collectively (humanity that is), just quit? What if we all stop playing? (I think you're saying this) -- But seriously. Both OpenAI and every other place (gawd I hope others aren't doing this) need to figure out their business model fast™ and make it a viable business without devolving into this tracking, marketing and advertising bullshit some of us have only seen far too often -- that have driven us away from things like Meta/Fac** and others...

What the hell is this?!

Why must every piece or technology, every software service or cloud "thing"™ devolve into Advertising?! That's it, I'm out. No more ChatGPT for me.

Someone should explain to Trump what a ZFLOP is (https://epoch.ai/data/ai-chip-users-documentation/methodolog...) -- Turns out if you build ever more powerful clusters of GPU(s) you can do some pretty interesting things (Super? Hmmm, maybe, but super intelligence?)

I'm not sure what you're doing differently, but I just created a Claude skill called /triage-alert -- All I have to do is basically call /triage-alert without any arguments, or if I want to, paste a specific alert that happens to be firing.

The Skill itself basically is a 6-step process and a simple contract with some guardrails:

(1) identify the alert, (2) size the problem, (3) find the signature in metrics and logs, (4) test the leading hypothesis against a second source until it is confirmed or dead, (5) write the finding + plan, (6) AskUserQuestion for the decision, (7) implement via the sibling skill and verify, (8) update memory if the cause was new.

There'a 3 scripts to automate some of the tasks so that they are deterministic and 3 data sources as references as well as MCP servers it has access to for metric, logs and alerting sources.

That's it. It not only diagnoses the problem every time, correctly, but also figures out what went wrong, and why, and proposes how to fix the underlying causes.

I then have a follow up skill called /incident-postmortem that writes up a full incident report and post-mortem for the next time to learn from and feed back into the system -- Which is basically Claude's own memory over time.

Wow! :O Finally, a legislator with enough balls to put up something that _might_ (just might) make corporations _actually_ care about security and privacy! I can't wait for this to start being adopted in other countries. It's about time!

LOL

Not quite. A token. bucket alone would not prevent a retry storm if you have a chain of services A -> B -> C -> D with a failure in D, you'd end up still having A, B & C all performing retries as they cascade through the chain, albeit (yes) at your configured rate-limits (using token bucket), but you'd still end up with an amplification effect.

So, effectively if A → B → C → D and D is failing, C may retry D, but B and A are discouraged from retrying the whole chain.

This is quite slever. I also really like the concept of an "Error Budget", inspired by SRE and SLO(s) no doubt :)

No details? Signin/Signup CTA(S), that's it?

I find it quite amusing that Microsoft has Copilot and now puts Copilot in literally *EVERYTHING* they can think of, then at the same time, they offer you access to Anthropic's own models through Copilot I guess MAI models (Phi?) aren't good enough?

First, OpenAI runs around screaming and yelling for OSS (and Chinese) models to be regulated and banned. Then Anthropic yells and screams the sky(net) is falling and going to kill us all, let's regulate and ensure AI has built in kill switches. And... Now Microsoft's turn. The rivalry is honestly becoming a joke. Can these big-tech corps grow the f*k up and play nicely in the sandpit?

I work in a highly regulated industry (banking), and I can tell you from first-hand experience that regulations like this that basically amount to 3rd-party licensed verification and auditing don't really stop bad things from happening, or the company at-large (especially the larger and more complex it becomes) from activities that years later end up as non-compliance issues, only for the company to get served a fine that is basically "the cost of doing business".

Not to mention, it would raise the cost of Software as a Service businesses as the cost of building and maintaining software goes up, as well as the technical expertise to maintain those systems. That would have a flow-on effect of reduced market as low-medium income folks are effectively priced out of the market, no long paying for "cloud services, and so on... It would be a disaster.