argee

They're just going to rationalize answers to this. You already know the real answer. FOMO, and fear of the Red Queen effect.

Can't you just look at memory bandwidth for generation?

Thanks for the reply. It would be great if you could introduce yourself somewhere in this comment section or in your HN profile, I cannot tell whether you intend to speak on Karthik's behalf.

If there was value to be had there, you would. It's an indictment of the cited value, if you wouldn't. So are you asking me to argue my own point again? I don't think your position is adversarial to mine?

If you are arguing effort versus value, let me escalate: I don't think even the call leader of an Amazon sev one would be interested in seeing such screen recordings, when all of Amazon.com is down, and when workable solutions would be worth putting Herculean effort behind.

I want to know where the value comes from, then, just because it got easier to query such data doesn't in itself make it valuable. You would be better served by having Opus look over your code after Sonnet made mistakes, rather than having Opus first digest whatever (potentially poisonous slop) Sonnet cooked up before looking at the code.

Would there have been value, pre-LLMs, in attaching a screen recording of what a developer did in the three hours before posting their diff (be it Google, stack overflow, reading a book, or trying a sandbox experiment) to their PR? The PR is a synthesis of their efforts, and it would actually be more effort to review and understand their screen recordings, wouldn't it? In my experience striving for fewer artifacts is the right move, so long as they are sufficient to describe the outcome. What exactly is lost between sessions and code that would help you debug something?

Also, if your business has agents writing most of the code, why would your business care about human reviews at all? In for a penny, in for a pound, after all. Wouldn't you be better off looping your agents into your support tickets and autofixing bugs? In that scenario, isn't it more confusing for an agent to deal with sessions + code rather than start fresh with the code?

Maybe your username could have been paulrichguy and things would've been different.

Are you conflating regulation and accountability? I haven't seen anyone (without a vested interest) ask for the former as a response to an adverse event, only the latter. There is a difference.

Not producing output at a speed you can't legitimately keep up with.

This reminds me the old story about the man who goes to the doctor complaining of pain in his eye when he drinks his tea/coffee, and the doctor advises the man to take the spoon out of the mug.

Except here, instead of taking the spoon out, the man just decides to take his eye out.

I suppose that does at least permanently solve the "problem" in one fell sweep.

This is just a form of "if you're so smart, why aren't you rich?", isn't it?

It's best to have a light hearted attitude about these things (that's why I added a joke). The vast majority of people don't even know their pronunciation might be different from others', ignorance is very human. It isn't malice, though, and even if people did make an effort they would be the odd one out.

One interesting example of a blanket change is the N-word. As expected, it wasn't a grassroots movement but a mass publicized event and collective experience that made it happen.

If people's ignorance makes you angry you're in for a miserable and likely hypocritical life.

That's the joke!

> In the Karakoram (pronounced keh·ruh·kaw·ruhm) Range—a mountainous region straddling the borders of Pakistan, India, Afghanistan, Tajikistan, and China…

Non-sequitur but no one pronounces it like that in those regions.

On another note katsu (pronounced cat-sue) is a dish you should try in Japan.

That's a funny example, given that the Wright brothers were overwhelmingly empirical compared to their competitors and succeeded as a result, because the foundation of aerodynamics was fundamentally flawed at the time. In a real sense, those systems were built from empirical observations, and most human systematizing happens post-hoc.

I agree that formalizing observations into a system can lead to additional insights, though it would probably be hubris to assume the formalization is faultless.

It doesn't sound great to me. That would be a bit like if I asked my contractor to read a 500 page tome instead when he wanted to come look at the studs in the walls, because the original architect made an incomprehensible mess of it.

I wouldn't put it exactly like this, but I agree that many systems chugging along today (such as the COBOL running on IBM metal that powers banks) are not something people try to (or can) comprehend all at once.

Is that really a problem though? We can solve problems empirically even if we don't keep the entire system's workings in our heads.

Wow, what an insightful comment, paired with a downvote. I am in awe.

Nevermind that the calculation of their retirement number was what I was suggesting in the first place.

It's amazing to me that even HN is full of this kind of petty jealousy, where people are ostensibly amiable to founding their own businesses and owning their time. God forbid someone retire. I expect to get downvoted on Reddit for even daring to mention I can afford bread, but I didn't expect that kind of thing here.

Maybe you should choose a different sin, you know Envy is the only one you can't have any fun at, right?

> The difference in an attention layer is that the query is matched to each key using a dot product, and rather than getting a single value back you get a weighted sum of values.

I understand you meant to compare it to dicts, but this is more like (the only?) similarity to outer product memory. Worth calling out since GP is confused about this connection specifically.

How much money do you need?! I don't find it a particularly hard sell to not sell off my time after having reached my retirement number.

That's true. What I really mean to say is, people with worthwhile things to say don't usually say them only on the internet, whereas people who didn't put any thought into what they are saying often do solely post their opinion on the internet, and that I prefer to look at higher effort output compared to internet posts as far as information is concerned. My point isn't about denouncing "people of the internet", but rather, the median quality of opinions on the internet. And the poem, of course, is about who is ultimately accountable.

Charlie Munger's famous rule is: "Never, ever, think about something else when you should be thinking about the power of incentives."

That rule has served me well enough!

Do you go to the circus and call the clowns dumb? Why even go?

Hmm, you'd have to go if you are yourself a clown, I guess.

These models couldn't code their way out of a paper bag and the frontier labs are still putting them in sandboxes made of thinly stretched cotton candy. It seems like their incentives are much more aligned with shutting down open source or Chinese opposition and not borne out of a "fear" of the models themselves.

> When Ledger hardware wallets came out and people said that seed (the list of words) could be intercepted and relayed by a channel bypassing the computer (here by using an eSIM), people would post explaining that people were nuts if they thought such an attack could be pulled off.

Every single time I thought maybe people on the internet know what they're talking about, I've been proven wrong. The internet is for entertainment purposes only.

The only educational exception has been people who weren't exactly posting on the internet, such as the Berkshire Hathaway shareholder meeting videos or Stanford lectures.

There's a lot of grief can come from it, so I try and read this poem often. [0]

[0] https://gladdestthing.com/poems/ed

The internet is increasingly full of the types of people it's pointless sharing anything nascent with anyway. I'm sure a local meetup will see your side project very differently than a subreddit or hacker news (the latter will even send armies of bots to take you down).

There might be some selection bias here, since using Claude code or Cursor et al can be done directly with first party API keys and those users never need openrouter.

> Like, why assume that your experience is necessarily the most authoritative, the most representative, when you encounter evidence that someone else's might be different?

Because it's easier.

That's awfully optimistic. Thinking, something humans have eschewed since the beginning of time, becoming something people do willingly, for fun? Haha.

Do you think about Nietzsche's Last Man?

> If we understood the models we could monitor them and make them behave how we wanted.

Charlie Munger's famous rule is: "Never, ever, think about something else when you should be thinking about the power of incentives."

So is it that we...can't? Or something else?

> If you say that we "understand the brain" because we understand how a neuron works, you'd be using "understand" in a way that mislead everyone reading that sentence.

What if we say we "understand the brain" because we can build one on demand?

> My comment was about what Billionaires do with their money and them lobbying for tax cuts, pushing down wages as far as possible or exploiting mechanisms to trap them into addiction is not "what people want".

Yes, and my comment was about the idea that these billionaires are often man-made, the population that funnels wealth to them could instead, in the first place, funnel wealth to the causes they ostensibly want the billionaires to support, cutting out the middleman. But very few people actually want world peace or a child in the third world getting a meal, as opposed to say, another can of coke or to stay home and get Doordash.

> This is usually how a prospering and successful civilisation works. You develop a justice system, laws and regulations so people don't smash their heads until their families are extinguished for example.

Agreed. It's par for the course and not necessarily evil. But worth being aware of the hypocrisy of deciding what "right" is, for others.

> @Edit I hope you saw the whole movie because dumping assets they know are nearly worthless on its own clients and counterparties before they find ou is not "giving people what they want". It works only because the buyers don’t know what they are buying.

Are you equating algorithms to fraud? That doesn't seem quite right.