Hacker Newsnew | past | comments | ask | show | jobs | submit | ibejoeb's commentslogin

Wow:

    In this role you will:
    - Design and build agentic AI systems that analyze, generate, and validate...
    - Build agentic architectures that compose specialized AI agents dynamically...
    - Build AI-driven continuous validation frameworks powered by agentic workflows and large language models that autonomously manage...

This is invoicing? If ever there was a domain that was purely deterministic, you'd hope it was invoicing.

When I was at AWS, they famously required an extensive "CoE", correction of errors, or post-mortem, in an instance of over-charging a customer $0.26.

The idea is that if we can make small billing mistakes like that, we can make large billing mistakes, and need to invest in the correctness of the systems powering billing.

I have great respect for the engineering culture within AWS during those times. I am glad to have left before seeing it degrade and decline.


The irony is, the only purely deterministic thing, will be token consumption...

I severely doubt the world ever gets to such a point that the entire world melts into AI hallucination. And token consumption depends on so many other things, it's not all that deterministic either.

(token usage) is trending towards predictability for a lot of reasons. it's not deterministic but it's getting easier to reason about usage.

How can a random generator be deterministic?


I’m not so sure about that. I can see a real rationale for creating sanity checks using AI to more quickly/proactively catch pathological billing issues before they become HN nightmare stories. They wouldn’t replace billing code, but there are many ways that stupid customer mistakes can cause real costs to Amazon that either have to be refunded and absorbed by Amazon or paid by the customer causing a negative opinion of AWS. If a billing AI watching costs in realtime could detect, say, a lambda loop in the first 10 min and either alert the customer or kill it, that would make AWS feel a lot safer to use. Enumerating these conditions and fixing them individually is a task that Amazon has proven incapable of achieving. An AI watchdog layer might be the perfect shortcut to addressing all of these problems at once. Because it’s well-trodden territory that AWS has so many multi-thousand dollar foot guns that make it really scary to use as a hobbyist or small business on a tight budget.

> I can see a real rationale for creating sanity checks using AI to more quickly/proactively catch pathological billing issues before they become HN nightmare stories

Right, so invoicing is still a deterministic problem. You can bolt whatever on but in the end it's just product x price x units


This is exactly the sort of thing that’s not possible, though. An AI will not be able to detect a “lambda loop” because it will look exactly like a “successful lambda rollout”. This sort of watchdog would just as likely shut down the wrong things and make AWS feel a lot less safe.

It doesn’t matter which domain it is, AI should be utilized at anomaly / patterns detection and prototyping, with goal of a prototype to become coded as discovered pathway.

I don’t understand why companies still place AI in their system at authority-forming level, if anything it should sit at least behind single guard / validator that is explicitly coded and tested, never in front of presentation layer.


This is like half of all job listings I've read recently. And it's a decent amount of fintech that's like this.

Probably not actually. Transferring one kilobyte across a network link has such a low value that the billing costs of aggregating it cost more than the revenue.

So instead you take a probabilistic approach - charge the user for a megabyte of data transfer 0.1% of the time, and bill nothing 99.9% of the time.

Now the typical cost is the same, the users bill is probably accurate to the cent, but you have divided the number of billing records by 1000.


The way you describe requires somehow counting every bit but somehow discarding most which is obviously nonsense.

This seems statistically invalid insofar as it will tend to overbill potentially by a lot on the minority of cases.

Don't you know how much of the pipe is occupied by a given customers code at any given time or what data is being sent


You have to do it when the customer list is too big to keep a counter per customer.

A probabilistic counter per customer is also a counter per customer. Still, probabilistic billing is an amusing thought though.

No you don't

I don't know how cloud services count usage, but this is certainly not true for telco. I manage several fleets of hundreds/thousands of SIM cards (mostly IoT/M2M applications), and almost every provider counts the data traffic per byte. Different business and use case, I know, but still.

Laws like that are just going to give rise to new tortured wording. You're buying a revocable license to view the content under certain conditions. It was already in that territory even with physical media; that's what region locking is. Likewise, if bitrot set in and your disc became unplayable, the distributor didn't send you a new one. You never had a perpetual, irrevocable, and otherwise unrestricted license to view that content.

I'm not saying that it is not worth trying to fix this, but now that the technology enables content owners to more fully control your access, they're not going to be keen to relax that only to leave that money on the table.


What? "You never had a license" - no, of course not, that's not what buying is. You had a phonograph record or whatever, and it didn't get replaced when worn out in the same way that the shoe manufacturer doesn't replace your worn out shoes which you have bought. Region locking, what about it? It's interference with ownership too.

Things you can buy have to be accurately described as what you actually get, so "buy this" ought to be an accurate description of what the deal actually is, too.


I get it, but your advertising to it what you want it to be, not what it is.

I don't have the answer, but I wonder if they are considered a utility and operate on utility easements. Then we have to look at the county and state, too.

On the other side, I've read they operate a considerable number of private installations, too. Even that is suspect, too, in that there is existing case law affirming that people have a reasonable expectation of privacy in "the whole of their public movements."


They’re not utility. But you don’t have to be a utility to construct in the right-of-way.

Good point. Then, if the contract lapses, who is responsible for that infrastructure? Does flock have to tear it down?

Probably some legal question that won’t get answered until there is a strong pressure to do so.

No, because they have other customers.

>LAPD can just call them and access the data

Can they? Does anyone know the terms of these contracts? Does flock just look the other way if a licensee just gives away the data to some other entity without getting a fee for it? I can see arguments on both sides from flock's perspective, i.e., revenue vs lock-in.


There are dozens of ways for it to be done, but here's an example of how Austin PD worked around contract termination:

https://www.kut.org/crime-justice/2026-02-12/austin-tx-apd-f...

The network effect value far outweighs the lost revenue from a single local contract.


Everyone is motivated. It seems to me that Anthropic went out fishing for something to blow up. As Ray puts it: "Management eagerly approved the Rust rewrite option because it was a great marketing opportunity to showcase their new Fable model, Anthropic already uses Rust, and Zig is openly against using Anthropic’s products." Sounds spot on.

Others have pointed to some plausible technical benefits from the port, namely that it forces an enumeration of unsafe code boundaries. There's probably something to that, but the big idea is that Anthropic bet that its models could do a big job, and they needed to both demonstrate that and make hay of it. Everything aligned. They did the port. The artisans might call it slop, but it runs, they did it like 2 weeks, and they stuck it to the guy that shuns AI.

If you're the type of manager who picks farming out work to a body shop with 100 unskilled workers instead of hiring 2 legitimate FTEs, then this is speaking your language.


none of this justifies a nasty mean spirited blog post.

I didn't say or imply that it did. I just don't believe that the decision to port bun was based on innate technical merits of zig vs rust. I also don't think that it is entirely meritless, as I mentioned, because now that there's a par rewrite, there is clear delineation of the areas that need to be ported to safe rust. Still, does anyone really think that this would have happened without Anthropic's provocation?

I see it like this: Anthropic saw an opportunity to make a splash. Jarred made it happen without directly disparaging zig, but the implication is that zig couldn't cut it. Andrew made it more personal.


I don't understand how a port is a "provocation." That's a very cynical and transactional way of thinking about open source.

(As for Anthropic's motivation: only they know. But I suspect that they saw Zig's AI policy as a potentially existential risk, and moved to de-risk their position. That is a technical decision, just not one rooted in the merits of any particular language.)


By provocation, I mean what motivated the port. I'm not suggesting that it was "provocative" like with intent to bother the zig team. That's why I ask if anyone thinks that bun would be ported to rust if not for the Anthropic deal. I doubt it, and I think that's why it took everyone by surprise.

Regarding zig becoming a risk: that's a reasonable line of thinking that I hadn't considered. I don't know how likely it is that zig will fail to progress and fall behind as a language compared to others that are assisted with AI, but it's a possibility.


Ah yeah, I agree with you.

I have a question: who trains the CSAM model?

That was terrible. But then the NSA just started surveilling everything illegally anyway, laws be damned.

I was just pointing out that the 4th amendment is not really helping the americans with tyranny and all the related stuff :)

There can be, but this isn't it. In the EU, a company can't send you an email, but it can read your email.

You should be using AIM OTR: https://otr.cypherpunks.ca/

Adding to this there is pidgin-otr and python-potr.

For IRC there is irssi-otr. There was weechat-otr but I think it may have gone unsupported as their script did not work in python3.

There is also ejabberd [0] that has OMEMO [1] preferred over OTR and PGP and supports many to many E2EE.

Someone tried to MitM Jabber, discussion here on HN [2].

[0] - https://ejabberd.im/

[1] - https://en.wikipedia.org/wiki/OMEMO

[2] - https://news.ycombinator.com/item?id=37955264


That's what they did.

> OpenAI invented a technique in July 2022 whereby its system would insert terms reflecting diversity (like “Black,” “female,” or “Asian”) into image-generation prompts in a way that was hidden from the user.

> Google’s Gemini system seems to do something similar, taking a user’s image-generation prompt (the instruction, such as “make a painting of the founding fathers”) and inserting terms for racial and gender diversity, such as “South Asian” or “non-binary” into the prompt

More links to primary sources, evidence, and official statements in the article at https://arstechnica.com/information-technology/2024/02/googl...


Lol fair enough, I stand corrected.

It seems then that their objective is to superficially increase the diversity of results, to avoid bad PR, rather than actually neutralising harmful and untruthful biases of the input data.


Consider applying for YC's Fall 2026 batch! Applications are open till July 27.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: