HNHacker News
TopNewBestAskShowJobs

jstummbillig

6,608 karma · joined February 14, 2015

submissionscomments
jstummbillig··on AWS Bedrock to require sharing data with Anthropic for Mythos and future models
That sounds like it's mostly just collective whitewashing, in face of essentially no guarantees when push comes to shove?
jstummbillig··on AWS Bedrock to require sharing data with Anthropic for Mythos and future models
Can you explain what AWS supposedly guarantees currently that your company values? I am not super familiar with the platform but I would assume, just like any other US company, that they will provide data to US agencies upon legal request as per CLOUD act, regardless of place of storage etc.
jstummbillig··on Claude Fable 5
In the "it works"* case: It's not even close. I did the math at some point (but I encourage you to talk it through with the LLM of your choice, there is obviously a lot of things to consider and weigh).

Anyhow, my research summary: Individual humans are so fucking expensive to train and upkeep (and this includes everything from before womb, where another human already limits their ability to work) You retain ~zero knowledge after death and start all over again for another measly 15 years of effective, productive work. Model training/r&d in relation, when deployed and used at scale, rounds to zero, even with the current retraining regime.

*Of course, the ratio can go to negative infinite if one assumes that models are doing 0 useful work currently and never will

jstummbillig··on Claude Fable 5
How do you arrive at that split? Real world is more like senior high level planning, implementation to juniors, review senior. Does this not translate?
jstummbillig··on AWS Bedrock to require sharing data with Anthropic for Mythos and future models
I suspect they will simply not offer it, for as long as they maintain that it has to in fact fly. Anthropic appears to be somewhat principled here.
jstummbillig··on Claude Fable 5
What subscription?
jstummbillig··on What it feels like to work with Mythos
I am sure you would not find it hard to exhaust any model, if you kept upping your ask enough times.

On the margins, suppose the prompt is literally: "Build a feature complete, high polish Facebook clone". Facebook is complex but likely not super complicated tech, and still I would assume that (after having burned through a substantial amount of tokens) you would find substantial enough differences in the outcomes between different models on that prompt on various fronts.

The above ask is obviously not useful, but what's preventing you from taking on bigger chunks until you approach the limit? At some point you would hit a boundary, where the diff will be obvious.

jstummbillig··on What it feels like to work with Mythos
Less fascinating when you consider that this is a non-coders perspective.
jstummbillig··on Claude Fable 5
I mean... if somebody gave you ONE prompt to audit a codebase, that might also burn 60% of your weekly usage. It's kind of a big ask, potentially.
jstummbillig··on Claude Fable 5
I wonder where the trees are. In this thread nobody appears to actually be talking about the model.
jstummbillig··on Claude Fable 5
But you can just say shit now. Tokens might not be too cheap to meter but saying shit increasingly is.
jstummbillig··on The iPhone's Last Stand
The anger is real, but it's misguided (as anger mostly is). The benchmark is reality. Everyone is more "TikTok addicted kids" than not and the analysis is quite apt.
jstummbillig··on Harness engineering: Leveraging Codex in an agent-first world
Well, to be fair, the amount of goalpost shifting that is going on is quite intense. AI not being able to work in a "serious" project, and being limited to "toy projects" has been a long standing critique.

But also, bigger projects need some amount of loc written and it's a bit silly to pretend that this is not the case or a bad thing.

So the answer to the question is roughly: Establishing that an agent can work in a large-ish code base is valuable, because 1) them not being able to do so has been a critique and 2) it's something that is required for a lot of software projects.

jstummbillig··on I design with Claude more than Figma now
Maybe a random employees mildly interesting blog post is not exactly where we need to look for a psyop. But that is maybe what they want you to think.
jstummbillig··on S&P 500 rejects SpaceX, also blocking entry for OpenAI and Anthropic
The financials will be very different when you start doing something entirely different, with all your might.
jstummbillig··on S&P 500 rejects SpaceX, also blocking entry for OpenAI and Anthropic
Let's say Alphabet shifts further to ~become a 100% AI company in the same way that Anthropic and OpenAI are. Should they be removed from the index? If not, why not?
jstummbillig··on S&P 500 rejects SpaceX, also blocking entry for OpenAI and Anthropic
> just that nobody knows

I don't understand. Guilty until proven innocent, because they... are too successful? What could possibly be the generalizable idea here?

Should we have a speed limit for too successful companies, even if they might be doing super valuable work? Who would we trust to be the judge of the potential havoc that bad capital allocation in such a moment might cause?

EDIT: To be more clear, I don't have any particular qualms with the S&P committee maintaining it's position. That part I find mostly interesting and goes towards the second paragraph.

The first one is reserved for the quote, which I do have qualms with. "Nobody knows" feels a bit weak when the implication, that someone could be doing something illegal, turns into a guiding principle.

jstummbillig··on If AI data centers are so great, why are they being built in secret?
That confuses how offense/defense works in software vs real world. Publishing your vault layout will not lead to a lot of people want to use the vault and in the process submitting "patches" that you then can use and make your vault better. A built vault is not something that is easily malleable.

More people can however work on a great plan to attack it, which only has to work once, to be worth it.

jstummbillig··on Changing how we develop Ladybird
> They've massively decreased the work they put in but still expect the same pre-ai reaction/gratitude when submitting a big PR.

I don't think that follows. They expect things to improve, they do something about it and might (unreasonably) be frustrated by what they think is a policy that stands in the way of quicker progress. The first part is certain, the second part less so, and the third is just speculation.

It's clear that open source project are struggling to understand what is going on and coming up with plans – like everyone else – but clearly there are better and worse ways to proceed in this new world, if popularity, adoption and progress are things you want to focus on.

jstummbillig··on If AI data centers are so great, why are they being built in secret?
The point is that there is zero positive correlation between a vault doing its job and telling everyone where exactly it is. The bank would happily have it hidden in different secret dimension, if they could. It makes the vault safer. It's better.

"Why would you not tell us how to get to the secret dimension if it's so great?" It's just a dumb question.

jstummbillig··on If AI data centers are so great, why are they being built in secret?
I mean the framing is so silly.

"If bank vaults are so great, why would you not advertise their locations". It's a mystery, is what it is.

jstummbillig··on AI outperforms law professors in Stanford Law study
Doing things that are well meaning, but ineffective is not great policy. The simplest alternative to doing things that don't work is always not doing them. Better ideas are of course welcome, but not required.
jstummbillig··on AI outperforms law professors in Stanford Law study
But does it really matter? It seems fairly obvious that AI is going to outperform professors. While the studies run, there are three more model releases that change the calculus entirely. I wonder how much we are learning with these studies about what is going on.
jstummbillig··on Larry Ellison: "Citizens will be on their best behavior because we’re recording"
Why?

Them not having your best interests at heart does not make them an adversary. Just like you not having the best interests of your grocer or work place or most other humans on the planet at heart does not make you an adversary.

jstummbillig··on Larry Ellison: "Citizens will be on their best behavior because we’re recording"
That is not the point though? I am not saying that there are no issues with big corps and their interests. Theirs not being yours is just not one of them.
jstummbillig··on Larry Ellison: "Citizens will be on their best behavior because we’re recording"
What a strange thing to say. Why would they? Do you have their interest at heart? Do you have the average persons interest at heart? Again: Why would you? That's by definition not how interests work. And that is roughly fine (albeit a kinder universe would be nice).
jstummbillig··on Michael Burry says neither SpaceX nor Anthropic is worth $1T
On average, all the time. By any reasonable measure, more wealth has been created the longer the time horizon. Looking at the S&P500 over time gives an indication.
jstummbillig··on Stop Ruining It
Eh.

> Or perhaps: Satisfaction in our work isn’t created by the boss. It’s what’s left if they don’t ruin it.

~no boss sets out to ruin employee satisfaction. It's a byproduct of having to integrate more realities into the smaller scope that employees usually care about, and that is just not easy.

Of course, most bosses are also not great at this – on average bosses are, like everything else, just average – but to assume that bosses "ruin" satisfaction and employees would be longterm fulfilled and create working companies if only left alone is polemic.

jstummbillig··on Michael Burry says neither SpaceX nor Anthropic is worth $1T
Minus, of course, all the times where it did not go bum, right?
jstummbillig··on The Pirate Bay Remains Resilient, 20 Years After the Raid
It's kind of cool but also kind of spooky: A legally targeted site just staying online for 20 years, not because of obscurity, but simply because it can't be taken down.

Are our institutions this incapable?

← PreviousPage 11 of 34Next →