aetherson

I'm honestly stunned that people here are performatively math-illiterate.

Some companies experience liquidity events. Therefor the value of equity in those companies is positive. Some companies go out of business. Therefor the value of equity in those companies is zero.

If N is the ante hoc chance that a company will experience a liquidity event, then:

N * X + (1 - N) * 0 = value of liquidity

X is positive. 0 < N < 1.

Therefor the value is positive.

The rate of having a liquidity event for startups, particularly startups past the A stage, is just obviously not ultra low. Like, is it better than 50%? Reader, it is not! But it's not sub 10%, either.

People can make intelligent decisions about equity without hyperbolically insisting that the chance that that equity will be worth money is one in a million.

I assure you that statistics do not show it is best to value equity at 0.

Founding dates? I don't know. IPO dates were 2007, 2019, 2023.

I've worked at 8 companies, 3 have resulted in value from equity (all granted pre-IPO), all of them in certainly much larger than "a decent holiday" level.

I think people have updated to be much too negative on the prospect of equity paying out. It's obviously much better than 1%, at least if you work at anything other than extremely early-stage companies.

The business model of Colossal is to patent and sell genetic editing techniques that they prove out in their deeextinction process, for whatever that is or isn't worth.

The uses that they cited to me were "additional pair of eyes in reviewing contracts," and, "deep research to get started on providing a detailed overview of a legal topic."

I know at least two attorneys who use LLMs productively.

As with all LLM usage right now, it's a tool and not fit for every purpose. But it has legit uses for some attorney tasks.

What the author is referring to there as GPT-5, GPT-5.5, and GPT-6 are, respectively, "The models that have a pre-training size 10x greater than, 100x greater than, and 1,000x greater than GPT-4.5." He's aware that what OpenAI is going to actually brand as GPT-5 is the router model that will just choose between which other models to actually use, but regards that as a sign that OpenAI agrees that "the model that is 10x the pre-training size of GPT-4.5" won't be that impressive.

It's slightly confusing terminology, but in fairness there is no agreed upon name for the next three orders of magnitude size-ups of pretraining. In any case, it's not the case that the author is confused about what OpenAI intends to brand GPT-5.

It's varied a lot in that time period. When people claim that "wages have been stagnant," they're usually referencing this dataset:

https://www.ibtimes.com/infographic-real-wages-production-no...

Which is "real average hourly earnings for production and nonsupervisory employees," and is the source of the graph that purports to show stagnation between 1973 and 2017 (it has in fact grown higher than 1973 since then).

Why that particular graph? Well, it excludes wages for "the bosses" and such, so perhaps it shows something more like the average person's experience. But also, I think it's cherry-picked because it shows a particularly grim view. For example, here's the real median household income chart instead:

https://fred.stlouisfed.org/series/MEHOINUSA672N

And you see that it's by no means stagnant.

If you want, you can find income by quintiles. In like the 90's and 2000's, the rise of inequality, top quintile earnings growth was much higher than bottom four. But that largely stopped in the 2010's and has actually reversed in the 2020's, with bottom quintile growth now the strongest.

Long story short: income is complicated, there are a lot of ways to dice everything. But if you look into the claim that "wages have stagnated," you will overwhelmingly see the "real average hourly earnings for production and nonsupervisory employees" data line. And regardless of how good or bad that particular choice of data is, it doesn't actually show stagnation -- it shows a big drop in the 70s and 80s, largely stagnant 90s, and then growth since then.

We've propagandized a generation of people into believing that there has been no wage growth for 40 years when the deal was actually that there were falling wages in the 80s, stagnant wages in the 90s, and growth since then (with a recession after 2008 that has been fully caught back up to since).

"In real terms" has never once been used to describe an individual's situation. Like if you see a chart that has "real dollars" on it, it doesn't mean, "the author of this chart got a massive inheritance and so their money became less meaningful to them."

I am willing to bet you any amount of money you want that 3-4 years from now, current-generation AI models will cost less in real terms than they do now, not more.

I mean, not to say that you might not have some explanatory power here, but the market is complex and difficult to untangle, and at least some analysts are predicting recession which will certainly have effects on Reddit even if it's not directly affected by tariffs. We can all cherry-pick individual stocks.

I don't think he is downplaying the effectiveness of what you can do with the current models. Rather, he's in a milieu (LessWrong), which is laser-focused on "transformative" AI, AGI, and ASI.

Current AI is clearly economically valuable, but if we freeze everything at the capabilities it has today it is also clearly not going to result in mass transformation of the economy from "basically being about humans working" to "humans are irrelevant to the economy." Lots of LW people believe that in the next 2-5 years humans will become irrelevant to the economy. He's arguing against that belief.

Note here that "complete success" doesn't seem to necessarily mean recovery of vision, just some kinds of testing of corneal health. The article vaguely says that "most" patients had some improvement of vision. I couldn't quickly find discussion of this in the actual study, but the study claims success in proving safety and suggests further investigation is necessary to show therapeutic efficacy.

Tripod was started by some people a few years ahead of me in my college and my social circle heavily overlapped with people who worked there, so lots of us who were still students did part time work there. Later on, I did a summer internship where I actually did software development, not just what we didn't yet call moderation.

One of my first part time jobs was as a "pornzapper" for Tripod, one of Geocities' competitors.

We had a tool called the "fleshfinder" that looked for accounts with lots of images and then lots of those images had lots of flesh tones, and would dump them into a moderation queue where we human pornzappers would glance over the images and then press a button to terminate their account or mark it as within terms of service.

I am old enough to have seen now 30 years of, "Maybe airships are coming back" articles.

#2 is fanciful.

Great. Happy to celebrate them too.

Your experience seems limited.

https://dev.mysql.com/doc/refman/8.4/en/partitioning-types.h...

Sure, any relational, ACID database is going to be hard to scale. But of the ones that people actually use a lot, Postgres is the hardest.

Hacker News tries to discourage rapid back-and-forth angry posting by making it impossible to reply to messages deep in the thread for a little while after they are created. If you wait a bit, you'll see options to reply to even deeply threaded messages.

I mean, these are all useful techniques for scaling Postgres, but, like... they're also a list of why scaling Postgres is hard.

Very proud of NYC in leading the way to having trash cans, an advance that will surely spread to other less developed cities nationwide.

Also don't have a stray cat problem....

Street gangs are really fascinating, and aren't, like, "well-oiled crime machines" in general.

Economically, you have to understand some of the function of a street gang as we currently understand them: it is to be able to provide retail drug sales without either elaborate vetting of clients (which many potential clients will be unable to comply with), and without having the police be able to trivially shut it down.

The thing that makes this work is the use of juveniles to take the money/hand off the drugs. The police can easily find and arrest them, but they don't suffer prison sentences or unsealed criminal records as a result. Adults monitor the process but do not handle the drugs or money out in the open, so it is non-trivial to arrest them.

As a result, gangs have to be attractive to juveniles (almost exclusively boys), and function in a lot of ways as a social group and/or a somewhat violent community sports team (the sport in question being "having fistfights with other gangs"). Adult criminals who lead the gang push towards more serious violence with economic goals (eg: seizing the most valuable territory) and less social activities, the teenagers who are in the gang push towards more recreational violence and more social activities. Depending on what the size of the economic opportunity is in the area, you get more or less professional-seeming gangs.

So for example you see all this kind of D&D-like creation of iconography and myths in street gangs, strange coded symbols that they use in graffiti and superstitious beliefs (Bloods avoiding the use of the letter "c" in their words because it evokes their rivals the Crips), that I, at least, felt an undeniable resonance with in terms of my own adolescence. This doesn't serve an economic rationale and I think that adult gang leaders mostly are annoyed with it as a distraction, but reflects the interests and proclivities of the rank-and-file members.

You also need most of the teenagers to age out of the gang and leave (the gang can't absorb lots of adult members), which means being somewhat relaxed about entry and exit to the gang.

Absent an economic rationale for the gang, like prior to the rise of organized drug sales, my understanding is that teenage-dominated gangs were in general just social circles for malcontent teenage males, who largely did, like, vandalism and antisocial pranks, not Real Big Serious Crime.

Prison gangs, lacking the juvenile component, seem to me to be seriously different, though I've never read much about them.

Adding PHP to your stack in order to drop a column from a result is a WILD decision.

Almost certainly the heat.