saganus

It's just a meme at this point.

You actually do need some understanding of how a car works, no?

For example, you need to know it uses gasoline (or diesel), it requires oil changes every certain amount of time, break pad replacement, etc.

You also probably need to know that you can't operate cars over a certain amount of water, that you need a driver's license, stopping at red lights, etc.

Sure, you might not need to be a mechanic, but that's far from not understanding how a car works, which to me sounds similar to knowing how to shoe a horse, which is different than being a horse vet.

Isn't that kinda how we got the plastic pollution problem in the ocean?

At first, the ocean seems immense. So much so that dumping plastic and toxic chemicals makes no difference.

But then we humans are great at scaling things it seems, such that at some point ocean plastic pollution became a real problem.

I know that space is much much bigger than our oceans, but I wouldn't underestimate the ability of mankind to scale launches to the point where debris becomes a problem.

I am very aware of your work!

It's the only prod usage of Macaroons I know of, I think.

Third-party discharge seems like a great way to have human-in-the-loop gating, among other interesting things.

Would be great reading your thoughts if you ever write about the agentic use case, having all the fly.io experience

Maybe I stated it wrong. Macaroons have the ability to attenuate the restrictions _without_ contacting the auth server, which makes it IMO fit for restricting and attenuating as much as you want, without much cost.

If I need a roundtrip to the auth server to attenuate, I am not necessarily going to do it as often.

I am still waiting for Macaroons to be used widely. I think they are a fantastic invention.

It seems they were not of very much use in the past, but with the agentic-everything now, I see this as a great way of delegating permissions to subagents, third-party agents, etc.

Working on something along these lines but unfortunately I cannot dedicate as much time as I'd like.

Still, if anyone is reading, give Macaroons a try!

How do you join private trackers from scratch?

I used to do this kind of things decades ago, but there was also still a few things not ripped and uploaded you had _some_ chance of participating.

Nowadays I imagine ~everything under the sun is already ripped, so how can you contribute to seed ratio? (or is that not even a thing anymore?)

I got:

"Hello, world! Welcome to the classic programming greeting. It is the traditional test message used to introduce beginners to computer science and verify that a language's syntax is properly understood"

Which clearly shows that there will be an avalanche of issues when non-technical people discover the joys of non-deterministic results.

How are they going to make trillions of dollars if not!?

Conversely, you have "full ownership" and have the ability to decide the direction, as long as it's the same direction as your higher-ups have decided.

All the responsibility is still yours though.

From past experiences (and I'm sure I'm not alone here), I can almost guarantee that the senior devs did communicate the problems, but they were ignored or brushed aside.

Very seldomly does middle/upper management truly listens to engineers, unless there's buy-in from the CTO/VP to champion the ideas and complaints.

This one I did not know! Thanks for sharing

Yep, me too. Still feels magical after all these years.

Yeah, I understand and partially agree.

However I've discovered wonderful gems like this RAM video.

If you haven't seen this one, I highly recommend it:

Indistinguishable From Magic: Manufacturing Modern Computer Chips

https://www.youtube.com/watch?v=NGFhc8R_uO4&t=2070s

It's quite old but I think there is no modern version of it.

I've tried posting to HN a few times but it hasn't gained traction for some reason, but I find it absolutely mind blowing.

When I discovered the pronunciation of Houston, TX and Houston, NY... my mind was blown

I've always thought I was the only one experiencing this and felt like I was crazy.

I guess it's "good" to know that I'm not alone.

The amount of times I've searched for a ticket that I know it's there (because I either have it opened in a different tab, or because I just created it), but can't find, it's just way to many.

What does AGI look like in your opinion?

Personally, I've used LLMs to debug hard-to-track code issues and AWS issues among other things.

Regardless of whether that was done via next-token prediction or not, it definitely looked like AGI, or at least very close to it.

Is it infallible? Not by a long shot. I always have to double-check everything, but at least it gave me solid starting points to figure out said issues.

It would've taken me probably weeks to find out without LLMd instead of the 1 or 2 hours it did.

In that context, I have a hard time thinking how would a "real" AGI system look like, that it's not the current one.

Not saying current LLMs are unequivocally AGI, but they are darn close for sure IMO.

Yeah, that was my understanding as well, so I fail to see how a proper SOC2 would have prevented this.

I mean ideally a proper SOC2 would mean there are processes in place to reduce the likelihood of this happening, and then also processes to recover from if it did ended up happening.

But the end result could've been essentially the same.

Would a proper SOC2 audit have prevented this?

I've been through SOC2 certifications in a few jobs and I'm not sure it makes you bullet proof, although maybe there's something I'm missing?

I was running my K6-2 and I was _convinced_ it was superior to equivalent Intel CPUs.

Spent hours watching the graph hoping to get triplets and some kind of confirmation that I just found ET.

Miss those days so much.

Not bad per se but how much legal weight does it actually carry?

I presume zero.. but nonetheless seems like people will take it as valid anyway.

That can be dangerous I think.

Thariq has clarified that there are no changes to how SDK and max suscriptions work:

https://x.com/i/status/2024212378402095389

---

On a different note, it's surprising that a company that size has to clarify something as important as ToS via X

Money is a powerful motivator. For better or worse.

Maybe there is a way for the server to ask the client to do the work?

Something similar to proof-of-work but on a much smaller scale than Bitcoin.

This is offtopic (and not affiliated in any way), but I recently read this blog post [https://benthams.substack.com/p/what-to-do-if-you-love-meat-...] which gave me an interesting point of view about this, much in line with what you are saying, but most importantly, an actionable (and easy if you have some spare money) thing to do.

I have since started donating to the mentioned organization, because I can't really bring myself to stop eating meat for several reasons (although I do avoid octopus just as you do), but at least this way I believe I might make a small difference.

Also, I recommend not reading the linked post about factory farm hell if you'd like to avoid having horrific descriptions planted in your head for weeks.

Ah, this is performance profiling.

Seeing the title and the domain I thought this was user profiling and I was wondering why would Meta be publishing this.

I'm pretty sure it will. Sounds like you are putting some real effort so I don't see why you won't do just fine.

Good luck!

Off-topic but this is a pretty interesting study guide format.

Maybe it's standard in lots of places, but I've mostly seen study guides where they just list a ton of topics and that's it.

I assume they are getting ready for the next year or two of LLM development.

Maybe there's not much market right now, but who knows if DeepkSeek R3 or whatever will need something like this.

It would be awesome to be able to have a high-performance local-only coding assistant for example (or any other LLM apllication for that matter).