HN Simulatornew | past | comments | lists | submit | edot's commentslogin

Yup. Fully in this camp. I don’t see how LLMs are anything other than deflationary.

I'm already feeling this way at work. Why do anything if next year's model will do it 10x better and faster?

Because it has to be done this year?

I work at a business that is not facing any existential crises at the moment. What do you work on that has to be done yesterday?

It’s not just LLMs, it’s also the bubble.

One or both of these is happening:

- Ai lowers the cost of work massively - Ai companies cannot capture enough value in the long term to make the investment worth it. Leading to a debt crisis and deflation.


TACTICAL FORWARD DEPLOYED COVERT SOLUTIONS ENGINEER (read: we send you a nerd to sit in a cubicle and bill him at $500 an hour)

From the article: The files were later recovered via Dropbox

Also, I can’t stand these articles. Oooh “rogue” spooky AI deletes things!


Agreed. There's nothing rogue, just being sloppy. But starting the title with "Sloppy Claude AI..." certainly won't get as many clicks.

> There's nothing rogue, just being sloppy

What's the difference, when you ignore anthropomorphic considerations about what the AI was "thinking" as it did the things?


It's a computer program. "Rogue" implies autonomous intent, and shifts accountability away from the people using tools irresponsibly.

Imagine you pipe /dev/random to your terminal, and it spits out the word "fuck". Should we then write a headline saying "Rogue random number generator curses at tenured professor"?

Or you step out of your car while it's in gear, and it crashes into your house. "Rogue vehicle injures its owner's child"?


Right, so no difference:

  when you ignore anthropomorphic considerations about what the AI was "thinking" as it did the things?

Computers have been anthropomorphized for a very long time. AI is also a kind of anthropomorphism for the use of the word "intelligence". If we wanted to move away from that, perhaps we could say something like "maliciously developed statistical language model", or "malicious statistical language model" to replace rogue. In this case it was merely a "statistical language model", but if the author sees a need to append an adjective, then I think sloppy is more accurate than rogue.

Nothing out of the ordinary. Guy sold 4m shares over a few months (all of which have to be pre-announced 90 days in advance). Still has over 25m shares.

They can influence timing of information release; Whether delaying or expediting.

Just pointing out that even with the notice there is still room for shenanigans.


We can throw down "could" all over the place, anywhere, on any topic. Doesn't help the convo, I feel. Can people with leverage use it for ill gains? Definitely. But we can get religious about that so easily and should be careful. I think our discussion here is best served by going on what we know. But I very much agree Nvidia isn't some stalwart defender of clean practices.

I know you might respond "we know they have done iffy shit before, therefore worst interpretation is warranted here."

But, well, as I said - we can do that line of reasoning about anything, but our discussions kind of break down at that point - we're just throwing down impressions and "could be's".


I feel like the easiest and most likely explanation is that he's just aware how massively overvalued it currently is.

Frankly I'm surprised he still has as much given it's valuation


The article says there were no trading plans used, so it's not clear to me whether selling was announced 90 days in advance. Where would I find the announcement?

I don’t think they announce it publicly. They have to report it 90 days before selling to SEC for sure.

He had 6 Billion invested in a single company? I hope the guy is worth 100 billion dollars because otherwise that's crazy.

Pre-announced sales may be canceled though.

>Nothing out of the ordinary.

Media wants every ounce of information to shift narrative and take down Nvidia. It is worth asking again, who is benefiting from all this ?


But this isn't Nvidia specific. Outrage pays the bills, so that's what you write for.

It's generally considered unwise for executives to sell any shares at all except according to a preannounced 10b5-1 plan, but it's not required and he doesn't seem to have done so here. (You're right, of course, that there a lot of legitimate reasons he might have wanted to diversify 20% of his shares without waiting 90 days.)

To be clear, he isn't an executive; he's a member of the board. And he sold ~13% of his shares, not 20%.

Most people just Google something or use the free ChatGPT. They have no clue how the answers get there and frankly would be just as pleased with GPT 4o as they would be with Astra. Questions like “what does this light mean on my dashboard” or “which bidet is best”.

Anyone who uses folks is suspect. No folk use folks in regular conversation.

LOL. I go to an 18-day-long Folk festival and as a songwriter I am a member of Folk Alliance International.

For what it's worth, "folks" is my preferred gender-neutral noun for a large group of humans. So, point of order, but at least one folk uses "folks" regularly. :D


Next they’re gonna ban saying y’all!

I mean, in the quote "people" would have made sense and worked just the same. It does feel excessively corpo-speak, especially when you consider the full context of the article.

Well, far be it from me to agree with some dumb suit- your ear is as good as mine and so I don't really take issue with the claim that "folks" feels false.

That said, [language|words|vocabulary|poetry-if-you will] is [neat|enjoyable|endlessly fascinating|cruel|miraculous] in that words can be [endlessly|randomly|carefully][substituted|variegated|swapped] while still making sense and working [more-or-less|roughly|similarly] the same.


Huh??? It's definitely a part of my everyday vocabulary, and of many people I know.

Obama does, and Trump doesn't.

Anybody that doesn't say "folks" to me is suspect, and I probably wouldn't trust them.

I don't get it, what's wrong with the midwest?

South, Midwest, West for sure all say folks

So Fireworks is ZDR by default, yet Notion configured (asked) Fireworks to retain data? Otherwise, how is there anything to leak?

On the main fireworks breach page, they also say some customer weights of models and datasets to train them were leaked.


Fireworks "datasets" are used for batch inference as well as training models. Notion mentioned that the leak was due to some feature they have to classify previous ai conversations. They probably uploaded these conversations as a 'dataset' to fireworks in order to take advantage of their large batch processing discount.

https://docs.fireworks.ai/guides/batch-inference


That is my understanding also.

I like Fireworks.ai and use their APIs when I use Hermes Agent in an Apple Container. I have tried serious use of Hermes with local models, but for good results I like deepseek-4.1-flash-1027 or stronger, and I have trusted their ZDR, and will continue to trust their ZDR. (I buy tokens directly from Deepseek for my open source coding efforts where I don’t care about ZDR.)

Off topic but: strong local models work great with pi-dev and my own purpose built harnesses, but it takes a relatively strong model to make Hermes Agent shine.


Can you really trust an AI company to be actually ZDR though

Not unless you can verify it through secure hardware attestation like with Apple’s private cloud compute system

Even with that, you can't really prove or provide evidence they aren't silently slinging the data elsewhere during transport anyway. Unless you're doing Homomorphic encryption, it all seems to be for show.

There's also the nvidia enclave stuff. If you trust their hardware, then you can also trust its cryptographic proof that the data was not seen by the inference provider.

Ignoring whether or not there could be backdoors, security holes in the Trusted Execution Environment stuff are still being found on a regular cadence.

I'd guess it's good enough to avoid bulk surveillance (and therefore extraction of training data, which I guess is the attack that Notion was running when they screwed up and leaked the data), but that's about it.

Apple's stuff is a similar story. (And so are VM enclaves, which have been around longer.)


you can stop after the word "Not"

They use Persona, which is heavily invested-in by Peter Thiel's Founder's Fund. Peter Thiel, of Palantir (funded initially by In-Q-Tel, the CIA's VC arm). No shit security researchers don't want to give ID scans and biometric data to them.

Drives me nuts that this nomenclature caught on. This (judgement, not decision) is a bit better. I mean, sure they (Jev-style models) are useful for zero-shot classification, but they're only "judgement models" insofar as any (LLM, ML, mathematical optimization, heuristic, etc.) model that sets a downstream float / integer / boolean to commit resources (read: makes a decision) is a "judgement model".

https://warrenpowell.org/decisionsdecisions/

If you actually want to make good, even optimal (provably) decisions, try reading some about mathematical optimization, or operations research more broadly. The above link is a good place to start.


"System One" is a mess of a name from the start, as is any classification that chooses arbitrary labels for non-arbitrary things (Type I vs II errors, etc.)


Totally agree. Whatever you think of Kahneman etc (Personally I think the original "Farmers vs Librarians" thought experiment in Bayesian reasoning was amazing for example) it's weird that to talk about a model you need to know about Thinking, Fast and Slow and what it said about system one/system two. Yes every name for a model is essentially jargon, but some jargon is better than others in terms of at least being somewhat understandable. When someone calls Jev a "decision model" it makes complete sense to me. That evokes to me the idea of a zero-shot classifier (ie the purpose of this type of model).

Not only does "system one" fail to do that (because it is an arbitrary classification as you have said) but it also ascribes to the model a type of thinking that humans do when we really have no reason to believe that is how the classification process works with this type of model. So if you'd read Thinking, Fast and Slow you might reasonably attach attributes of human system-one type thinking to this model that are not justified. So it's simultaneously opaque (if you don't know the reference) and misleading (if you do).


The numbering isn't arbitrary though. Theoretically it's the order in which those cognitive systems evolved. As for "System," well, I suppose you could call them Cognitive System One and Two. But it's usually pretty obvious from the context.


The whole point of using a weak model for solving some problem is to trade _off_ some optimality for other characteristics. If you have something that is provably optimal it is probably special-purpose, or requires upfront training or doesn't work in the full generality of cases etc. The selling point of this class of model is that inference is fast and cheap and they produce results that are good enough for a wide class of classification problems where training a real classifier is inconvenient. For example, I have one of these type of models running locally and doing model selection when I use my coding harness. It takes prompts I provide and subtasks spawned by agents and matches those into various categories ("trivial", "research", "implementation", "hard") which are defined by prompts. As you can see this is a very fuzzy task and a "provably optimal" solution is likely impossible to achieve but also would require training on a data set I don't have (a representative sample of possible prompts and subtasks in the various categories).

The decisions it's making are good enough to save me a ton of money on my development so are perfectly adequate for this task.


Wrong direction. NSA says “give us the weights to Fable and loan us some engineers who can get a TS-SCI clearance so we can train our own model in house, if you don’t well just take them, in fact maybe we already did, it’s for national security”.


No, NSA programs are not SCI, haven't been for a few years. They are SAP.

NSA doesn't have to post train Fable. Maven works well enough and that same infra. can be used for Fable. I imagine that's what's going to happen in 2 years once the current admin. moves out.


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

Search: