keenmaster

Thanks for pointing that out, they’re actually tied for second-to-last.

As for % tested, states that don’t mandate the ACT tend to have higher performance in general. They don’t have as compelling of a need for the mandate, and they have many students who’d rather just take the SAT on its own. There is an effect going the other way though - if you don’t mandate the ACT, then students who don’t want to take any standardised testing at all…won’t, and so they won’t depress the average score.

The main changes to reading instruction were made around 2013, and math instruction in 2016. I wouldn’t expect a decrease in ACT scores 12 and 9 years later respectively, even when considering that it takes some time for instructors to master the new approach.

Mississippi's average ACT scores are tied for last (edit: tied for second to last). I’m sure some of their educational outcomes are improving, but the demographic-adjusted stats from elementary school students are misleading. Holding kids back for poor performance can really pump their numbers inadvertently. Even if that’s not a very prevalent practice, performance in high school is more important and far more predictive of life outcomes.

You know what’s crazier? Mississippi’s average ACT was higher before some of their education policy improvements.

The market ultimately determines the discount, not the company. In a competitive market, some of the gains from Apple’s change will accrue to the consumer, and some will accrue to the developer. What % goes to whom depends on demand elasticity.

Even if a particular list price doesn’t change, I’d expect more frequent and deeper sales.

In a less competitive market for a good or service (due to lack of antitrust enforcement) there should still be discounts, in proportion to the residual competitiveness. E.g. the mobile game market is very competitive, so I’d expect more discounts vs. the video entertainment market where there has been a lot of aggregation.

For Plus users the limits are:

o3-mini-high: 50 messages per week (just like o1, but it seems like these are non-shared limits, so you can have 50 messages per week with o1, run out, and still have 50 messages with o3-mini-high to use)

o3-mini: 150 messages per day

Source for the latter is their press release. They were more vague about o3-mini-high, but people have already tested its limits just by using it, and got the pop-up for 25 messages left after sending 25 messages.

It's nice not to worry about running out of o1 messages now and have a faster model that's mostly as good (potentially better in some areas?). OpenAI really needs to release a middle tier for 30 to $40 though that has the same models as Pro but without infinite usage. I hate not having the smartest model and I don't want to pay $200; there's probably a middle ground where they can make as much or more money from me on a subscription tier that gives limited access to o1-pro.

I don’t think early AGI will break out of its box in that way. It may not have enough innate motivation to do so.

The first “break out” AGI will likely be released into the wild on purpose by a programmer who equates AGI with humans ideologically.

If the first AGI is a very uneconomical system with human intelligence but knowledge of literally everything and the capability to work 24/7, then it is not human equivalent.

It will have human intelligence, superhuman knowledge, superhuman stamina, and complete devotion to the task at hand.

We really need to start building those nuclear power plants. Many of them.

AGI will commoditize the skills of the owning class. To some extent it will also commoditize entire classes of productive capital that previously required well-run corporations to operate. Solve for the equilibrium.

You have so much time to figure things out. The average person in this thread is probably 1.5-2x your age. I wouldn’t stress too much. AI is an amazing tool. Just use it to make hay while the sun shines, and if it puts you out of work and automates away all other alternatives, then you’ll be witnessing the greatest economic shift in human history. Productivity will become easier than ever, before it becomes automatic and boundless. I’m not cynical enough to believe the average person won’t benefit, much less educated people in STEM like you.

The nice thing about capitalism is that it doesn’t care what a company wants (in a properly functioning non-monopolistic market. I’d say that the international market for automobiles is more functional than not, despite some degree of subsidisation and tariffs). I’m sure Ford wishes it could charge you $100k for a Ford Explorer but it can’t because it’s not part of an oligopolistic centralized pricing mafia.

If self-driving tech becomes cheap enough, and there’s enough demand for it, then competitive pressures will ensure that people can own self-driving cars. That being said, the same competitive forces can make using a fleet much cheaper than owning a car, and you might want a subscription to the Mercedes fleet rather than owning an autonomous mid-tier car if those things are equivalently priced and you live in a fleet-dense area.

You are correct, and there are genetic studies to back you up. Modern day Egyptians have substantial continuity from ancient Egypt. Even in Ancient Egypt, there was trade and mixture with people from the Levant, but that didn’t massively change Egyptian genetics.

As for your question, you probably suspect the answer - many people will discredit Middle Easterners (including North Africans if you consider them distinct), past and present, inadvertently or not, intentionally or unintentionally. They are the modern day scapegoats, and nothing good can come from scapegoats right?

The biggest boogeyman in particular is the Arab. Muslim Egyptians, which are the vast majority of Egyptians, have a minority % of Arab ancestry. Oh God forbid, people of a shared faith but different ethnicity occasionally intermix. /s

Sometimes contentment is followed by a good partner, not caused by one. To reach that state of contentment usually requires some mix of effort and satisfaction with the outcomes of that effort and ultimately your confidence in your ability to maintain and grow the things you value in life. If you worked hard but the outcome isn’t good - why? If you achieved a good outcome (like a solid degree, a good job, and/or home ownership) but are particularly unsatisfied - why? If you’re not confident in your future - why? It’s not a requirement to settle these questions before finding companionship, but it helps.

No doubt a good partner multiplies contentment, but they shouldn’t be facing a void of it either. Lastly, if you find that your contentment is being pulled and pushed and stomped on by external factors, then do what is needed to gain control of it again - anyone can achieve this, and if someone in a land of possibility thinks they’re the exception, that the world really is getting in their way, they’re holding themselves back with that very thought.

Often in America the government waits for something to fail miserably before engaging in a high effort high cost activity that requires a lot of coordination and public buy-in.

I assume the two would be discernible from each other in your brain data to a sufficiently advanced AI. One is your visual stream and one is your visual imagination. On top of that would be signatures of intention to deceive, which are also detectable in theory :)

Florence is absolutely not a tourist trap. It is a beautiful city that was one of the wealthiest republics in Renaissance Italy. Great art (Uffizi Gallery is amazing) and architecture (i.e. Duomo) abounds. It is an essential visit in Italy alongside Rome in my opinion, though Rome deserves more time.

That just depends on where your focus is. If the safety issue is much more salient, you’re not as prone to recognize the humorous aspect of the situation. If you see the image on a meme page, you’re already primed for comedy.

Better magnets can improve MRI resolution eventually

Bing Chat is different. It can pull directly from internet sources, including article previews.

More competition increases the number of jobs

Ars tried and failed to reproduce the text, and then assumed that OpenAI closed a loophole when in fact they may have changed nothing - it’s just that NYT hired a team of experts to make tens of thousands of attempts to break the normal behavior of the model, while Ars didn’t:

“ ChatGPT has apparently closed that loophole in between the preparation of that suit and the present. We entered some of the prompts shown in the suit, and were advised "I recommend checking The New York Times website or other reputable sources," although we can't rule out that context provided prior to that prompt could produce copyrighted material.”

If I tried to do what NYT did, I would:

- give up after a few attempts

- get worried that I’m going to be banned from using OpenAI (a rational concern, because it doesn’t take a lot to get banned)

- doubt the veracity of any “reproduced” text that I may receive

The equilibrium here is for OpenAI to suggest signing up for NYT and providing a signup link if it detects interest in doing so.

It’s beating loose slander with less loose slander. Seems fair game to me:

“ OpenAI believes that it took tens of thousands of attempts to get ChatGPT to produce the controversial output that’s the basis of this lawsuit. This is not how normal people interact with its service, it notes.”

I think the substance of OpenAI’s complaint is valid - think of the word “hacking” as clickbait to the real heart of the matter which is that New York Times went to obscene lengths to reproduce meaningful amounts of article text and, by withholding their methodology, misrepresented the behavior of OpenAI’s product. There are much easier ways to get NYT content for free than making tens of thousands of attempts to reproduce an NYT article while violating OpenAI terms of service and repeatedly ignoring GPT’s refusals to an insane* degree.

*not normal insane, insane to the 10th power

Don’t underestimate the elasticity of supply.

Genuine question: you don’t think that outlier analysis (with moderate diligence) would have picked up the error in the original study, despite the plotting library issue? Just look at this link - this wasn’t detectable with any type of cross-validation? https://arxiv.org/html/2402.14583v1/x1.png

In my opinion, this thinking is wholesale incorrect and actually terrifies me more. Yes, there is global warming. Yes, it is extremely concerning, and leads to countless endangerment, extinction, and coastal/island loss etc…it will cost untold billions to mitigate its effects in even a handful of cities like Miami, Venice, etc…

But, exactly why should people have fewer kids? It’s the large, sweeping, gubernatorial and industrial changes and innovations that drive meaningful shifts in sustainability (with downstream effects on consumer sustainability i.e. cheaper EVs heat pumps and solar panels = greener consumer behavior). If we have fewer kids, that’ll just add to our problems - we’ll have climate change and accelerated economic collapse. If you survey people from developing countries on the anti-natal perspective, who emit a fraction of the C02 that people in developed countries do, they might find it borderline pathological. Malthus was dead wrong, and his intellectual successors continue to be wrong, just with more cope. I have no reason to believe that will change.

Anyway, people who think that having fewer kids is as much of a disaster as climate change will continue to have kids. People who think that they should go against their core reproductive instinct in order to save the planet (???) will not have successors, if their behavior is consistent with their ideology. It is only the children of the former who’ll be able to evaluate the outcome. I assume they’ll be generally thankful that they’re alive to evaluate anything in the first place. I also assume that they won’t be living in a literal hellscape - point me to any credible literature which portends that for people living in the United States in 2100 (because that’s where most people on HN live - there are people today that live in a hellscape, they don’t even have to wait for 2100, but it’s not predominately due to overpopulation).

Bizarre. How do people make such big, splashy findings that can mess with people’s sense of optimism about science and innovation, without doing the simplest types of checks on their data and methodology.

MicroLEDs will fuel the next display revolution and are smaller than some bacteria. They can eventually achieve the perfect pixel density for AR/VR while having better display characteristics than even OLEDs.

Why aren’t the big models trained on synthetic datasets now? What’s the bottleneck? And how do you avoid amplifying the weaknesses of LLMs when you train on LLM output vs. novel material from the comparatively very intelligent members of the human species. Would be interesting to see your take on this.

Exactly. That’s also why I find low parameter LLMs to be useless for me personally. I simply cannot trust anything that is so very illogical. GPT-4 is the first LLM that crossed into usable territory for me. Even GPT-3.5 was a fun toy and maybe good for summarization, but that’s it. It will be revolutionary when GPT-4 is cheap enough that thousands of calls don’t cost much. To imagine an LLM much smarter GPT-4…the future is bright.

They’re used in political campaigns. Do I need to say more?

It’s all relative. They’re LCOL on a per square foot basis, if you have a family or otherwise need more space.