polygamous_bat

However if you are playing for the regulatory capture route (which Sam Altman seems to be angling for) it’s much easier if your name is “OpenAI”.

When the time for making money comes, if you don’t think OpenAI will sell every drop of information they have on you, then you are incredibly naive. Why would they leave money on the table when everyone else has been doing it for forever without any adverse effects?

I’m a little out of the loop, could you please elaborate?

I think it’s less of a 2D structure and more of a tree structure that you are describing. I’ve also felt the need of having “threads” with ChatGPT that I wish I could follow.

Hey stop shaming the GPU poor, not everyone is Mark Zuckerberg ordering $8bn. of GPUs.

“Government” isn’t a magical boogeyman coming out of a magical swamp to make magical laws. It’s literally NIMBYs, which are largely current property holders which are also in large part landlords. They’re playing the game to the full extent, and it’s high time they are contested in the same game.

Why would you rent for 50K if you can rent for 100K? Are you an economically irrational actor?

Same applies for landlord who could either undercut the competition or join the “app” (aka collusion ring) to make higher profit.

> With housing it's more flexible: you can commute longer each day. You can just not move to a city where you know the "salary/cost of living" is not working in your favor. You can also time these decisions: go to NYC when you are in your 20s, live with some roommates to get some savings and then when you get older you can move to a more spacious city. You can be flexible about the space you need too: from 1 bedroom in a shared apartment, to a 3 beds detached house in a quiet suburb in another city, there is a lot of flexibility.

You _really_ think people aren’t already doing that, huh. Rents are too damn high AFTER all these options have been exhausted, and not because (as Fox News will happily tell you) people are lazy and entitled.

So our choice is to either legally discourage rent seeking behavior in certain necessary parts of our life (health, housing, education) OR “live in a van… don’t go to a doctor … go to community college” just so the overlords could keep making profits on their investments. I wonder why not more people are signing up for the latter.

Think logically for one more second. Very similar logic applies for health insurance/cost of college/rent, because those are “unavoidable” in some ways if you want to live (with dignity, because homeless people are still alive I guess).

The only reason health insurance/college/rent costs near infinite dollars rather than actual infinite dollars is because price hikes work best when they are hiked slowly every year instead of all at once. Health insurers/bloated college admins/landlords would (rationally) love to have infinite dollars, as they have a stranglehold on a necessary resource, except that’s probably a quick path to the guillotines coming out.

Market determines rent the same way market determines the cost of health insurance: people need a place to live the same way they need a doctor when they’re sick. If all the landlords are using collusion software to raise rent in tandem [0] what is your alternative, live in the streets until they give in?

[0] https://www.propublica.org/article/yieldstar-rent-increase-r...

> 10 seconds is absurdly slow, though.

I hate to say this, but have you used any $ModenWebApp with $HotJSFramework recently? I thank the gods when those pages load without a 5-10 second of fancy spinning animation. Really thought we would be in a better place by 2024 but nope.

I think the flawed assumption there is thinking management is there for the good of the company. In my experience, much of the time management is there for the good of itself, and when it’s good for the company it’s mostly a byproduct. Just like how we use yeast to make bread and beer.

Automatic != hidden. Painless != hidden.

The fact that you want to cause extra pain to those who can't pay big money lawyers to simply make a point says more about your political leanings than anything else in this comment.

For those unaware, OpenAI recently announced [0] an API change where they said their newer models are using Matryoshka representation learning for shortening embeddings. Basically you can use a shorter prefix of the full representation to do query/lookup for cheaper without losing much quality. Quote:

“Native support for shortening embeddings:

Using larger embeddings, for example storing them in a vector store for retrieval, generally costs more and consumes more compute, memory and storage than using smaller embeddings. Both of our new embedding models were trained with a technique [Matryoshka Representation Learning] that allows developers to trade-off performance and cost of using embeddings. Specifically, developers can shorten embeddings (i.e. remove some numbers from the end of the sequence) without the embedding losing its concept-representing properties by passing in the dimensions API parameter. For example, on the MTEB benchmark, a text-embedding-3-large embedding can be shortened to a size of 256 while still outperforming an unshortened text-embedding-ada-002 embedding with a size of 1536.”

[0] https://openai.com/blog/new-embedding-models-and-api-updates

This is the same "I can fix twitter search in an internship" George Hotz, who then gave up and ran away before half of his internship was over?

When would we be ready to admit that being a world-leading expert in one area (jailbreaking devices) may not transfer to others (like urban design) if it doesn't even transfer within the domain of CS? The sooner we can stop worshipping the ground these people with an expertise on one thing but opinions on everything walk on, the better off we will be.

Because otherwise when you sell stocks you get prosecuted by the SEC.

Not to be pedantic, but I imagine there would be easier ways to telling a Japanese soldier apart from British/American soldiers during WWII /s

Immutability is great for multithreaded/async programs because every thread can rest assured knowing no other thread can sneakily modify objects that they are operating on currently.

I haven’t checked the proof yet, but could it be possible that it’s using directed angles? Threw me off initially when I first learned it as well, but it can be handy for different configurations.

I assume OP means social skill in the sense of “skill with society as the instrument/playground” rather than “skills that benefit social bond”. Same way as a 10x dev at a nonprofit and a Russian hacker can both have “computer skills” while doing vastly different things with their skills.

As a current AI researchers, my opinion is while AI may surpass the average human soon (if not already) in a game of 9s humans are always better. So I wouldn’t trust a system with 0 humans in it for failsafe.

Ah I am sorry, seems like a comment was posted later (Jan 9th) providing a justification of their criteria. However, the reply comment it was posted on wasn't there when I first read the article and the comments on Jan 8th.

> Moreover, saying "no one knows what happened to those 11,285 studies" without checking in with the authors is completely unfair. The first author responded with the code showing exactly how they achieved that figure. Nothing mysterious.

But that is the whole point! The methodology of how they dropped the 11,285 studies was not even told in the original paper, and even in the comments the author doesn't explain "why", just "how". Hence, I think it's completely fair to call it "irreproducible".

The point of doing "reproducible science" is not that I write a paper, you email me asking how did I come up with my number, and I email you back an explanation. No! The important details should be in the paper already. You may do some magic on your dataset, and that's fair _as long as you detail what magic you did and why_, and given you can defend that practice in front of your peers. Otherwise what is the point of preaching "reproducible science" at all?

Pretty cool, but please put your screenshot on the GitHub repo as well.

Wouldn’t someone think of the tax-evading billionaires? After all, that could have been you or me if not for a couple twists of fate /s

Because unlike what “Fountainhead” by Ayn Rand says, checks and balances on positions of power is usually a socially good thing.

Because some people are social and have fun with other people.

> The “let’s make life shittier for people so they’ll do the things I think are right” is bad politics and bad policy.

So do you propose we lift all bans on drunk driving as to make life slightly more enjoyable for a subset of society?

At some level all laws trade off “making life shittier” for some to “make life better” for society. You can argue against particular laws but categorically being against it reduces your argument as asking for anarchism.

> Ethereum really is the most impressive open source project in the world, IMO, because it's a live system with half a dozen different clients written in different languages that all have to interoperate with each other and perform complex cryptography in short timeframes.

I am curious, are you aware of linux?