indigodaddy

“Generative AI is intentionally not deterministic, and that means you're going to solve the same issues -- hallucinations, primarily -- over and over again.”

I see it differently. Sure there are still some hallucinations (but less and less as we progress on), but I see the AI-is-not-deterministic characteristic as a good thing, in the same way that humans and human designers not being deterministic is a good thing. Imagine giving a project/task to a human, and it was completed exactly in the same way with exactly the same outputs as another human that the project was tasked to. Would that be better than each human having varying outcomes of the project? I think most would agree the answer is no.

Unfortunately, the post comes off as AI-written. Why not just write your own posts?

Didn't know about this doc, thank you.

Yike, this notion is deeply troubling to think about

What we need is an AI gateway/router that will actually first analyze the input tokens and then decide what model to use. If it's so simple that a dirt cheap qwen 3.5 flash or whatever will be fine, then it chooses that. If it deems we need GPT 5.6, then it uses that, etc. does anything like this already exist?

So do/can users get a preview of the content before being required to make payment? Or is the idea that users are likely already familiar with the site/content to which they will pay? Sorry I only skimmed the article.

Or you could just use a great agent/harness like Shelley that already uses sqlite to store converians is searchable from the UI as well.

https://github.com/boldsoftware/shelley

Q: If/when Fable decides to nerd down to Opus on requests it deems dangerous, will we still pay the Fable API token rate?

Ignorance? I did say "competent" didn't I?

Did most of the companies have external facing jump servers? I'd hope at least companies have internal-only and even then with strict internal network access policies (+VPN etc) and ldap authorization etc. Can't imagine that any competent orgs would have externally-facing ssh or windows bastion hosts.

Yeah crawshaw, you should check that out... ;)

You basically described exe.dev

I'm interested in using these images on exe.dev. exe supports any oci images and stands it up as a microvm, in which it would be used non-ephemerally from that point. I'm assuming the images don't have any rc/services. How hard would it be to pull that back in after image deployment? (Also looks like I'd want to use the -dev images which include shell/apk, etc)

Apparently, NOT a theonion article

And..... I was right to be nervous

Lol I get nervous when I see a list of products with full specs but no prices

Just use Plexus [1]. The maintainer is not trying to be a hero or raise seed dollars or even really trying to promote it. He's just making an excellent, useful product. (Unaffiliated, just a happy user). It's not a full-on "LLMOps" platform (whatever that is), it's just a proxy that works very well and has some nice features.

[1] https://github.com/mcowger/plexus

Only non-pro is multimodal I believe

Kinda sorta analogous to the cloud engineers who can standup complex monstrosities in AWS-land, but don’t know the first thing about how to troubleshoot say a connectivity or simple problem where they have to ssh to an ec2 box and do the needful

Up to 50 VMs with your shared resource pool for $20/mo, just all their unique features like http integrations and Shelley of course is just native, their exeuntu image has everything you need. Build with Shelley, no need to deploy it's already on the VM (tell Shelley to setup systemd to "productionize" it) with a my-vm.exe.xyz url with auto-https via exe's https proxy where you can keep it private to your exe login or make it public (no auth through the exe https proxy).

I'd say try it out you won't be disappointed.

I'm a bit of a fanboy, but exe.dev + their Shelley web agent is pretty great

Pretty sure you could accomplish this in a large physical server or even a huge resource VM (that has KVM passthrough) with some sort of microvm technology? Then that would obviate the need for "multiple cloud instance per coding thread", it would just be a microvm on the large server.

Then again, I'm just the guy running his mouth, and you guys are the ones actually doing the work :)

BTW, looks very polished and thought-through, I may have to still give it a try!

I might use this if it supported any old cloud or VPS, and was at most $10/mo. The fact that you have decided that this platform should only live in your own custom cloud is unappealing to me.

Or, open source it and let us run it on our own VPS and keep your expensive cloud for those who want to pay. As it stands would never consider it.

My project is kinda sorta similar but uses Incus (LXC) and Caddy: https://github.com/jgbrwn/vibebin

Harkens me back to Sun/Sparc hardware and Solaris.

I've heard non-Pro isn't nearly as good for coding as Pro?

Would humans be cheaper?

OVH has FreeBSD templates. And most KVM VM/VPS providers will allow console access and mount custom ISO to install whatever you want.

No one will ever do this, or definitely not enough people will, so what's Plan B?

Huh?