no_op

Even if AI advances continue, for quite a while there's likely still going to be the 'Steve Jobs' role. That is, even if AI coding agents can, in the future, replace entire teams of SWEs, competently making all implementation decisions with no guidance from a tech-savvy human, the best software will likely still involve a human deciding what should be built and being very picky about how, exactly, it should externally behave.

I don't know if it makes sense to call that person an SWE, and some people currently employed as SWEs either won't be good at this or aren't interested in doing it. But the existing pool of SWEs is probably the largest concentration of people who'll end up doing this job, because it's the largest concentration of people who've thought a lot about, and developed taste with respect to, how software should work.

The author posted new results using the API (apparently the original run was through Codex), and 5.5 moves to the top: https://x.com/VictorTaelin/status/2047818978664268071

It's got a phone SoC. The use case for this thing is stuff you could do on a phone, but for which you want a larger screen and/or a keyboard. Web browsing, writing a paper for school, household budget spreadsheets. 8 GB is still basically fine for this.

A lot of this could be standardized and packaged into a product, a modern take on the 'server appliance.' Unpack some gear, plug it together according to a nice diagram, connect to a management console that feels familiar to anyone who's deployed to the cloud.

Apple's "strategic vision" for AI is to add a computer use agent (AI assistant) to the OS to perform tasks on behalf of the user, plus contextually surface AI capabilities in many specific contexts they've got utility (copy editing, image generation, photo organization, translation, coding).

What's missing here? What else should they be doing? What are their competitors doing, in any space relevant to their markets, that's much different? None of these critiques ever seem to say.

If AI ends up being another 'normal' technology, Apple's advantages in distribution (~2B active devices, with a user base that installs updates pretty reliably), ability to give their AI tools access to your existing data and apps, and general facility with packaging tech so consumers actually understand what it's good for, put them in an extremely strong position to capture value from it.

If AI ends up being something other than a 'normal' technology, if we really are a few years from building the sand god, well, all bets are off, and it's a little silly to evaluate the strategic planning of an individual company against that backdrop.

The material finding of this paper is that reasoning models are better than non-reasoning models at solving puzzles of intermediate complexity (where that's defined, essentially, by how many steps are required), but that performance collapses past a certain threshold. This threshold differs for different puzzle types. It occurs even if a model is explicitly supplied with an algorithm it can use to solve the puzzle, and it's not a consequence of limited context window size.

The authors speculate that this pattern is a consequence of reasoning models actually solving these puzzles by way of pattern-matching to training data, which covers some puzzles at greater depth than others.

Great. That's one possible explanation. How might you support it?

- You could systematically examine the training data, to see if less representation of a puzzle type there reliably correlates with worse LLM performance.

- You could test how successfully LLMs can play novel games that have no representation in the training data, given instructions.

- Ultimately, using mechanistic interpretability techniques, you could look at what's actually going on inside a reasoning model.

This paper, however, doesn't attempt any of these. People are getting way out ahead of the evidence in accepting its speculation as fact.

I think Moravec's Paradox is often misapplied when considering LLMs vs. robotics. It's true that formal reasoning over unambiguous problem representations is easy and computationally cheap. Lisp machines were already doing this sort of thing in the '70s. But the kind of commonsense reasoning over ambiguous natural language that LLMs can do is not easy or computationally cheap. Many early AI researchers thought it would be — that it would just require a bit of elaboration on the formal reasoning stuff — but this was totally wrong.

So, it doesn't make sense to say that what LLMs do is Moravec-easy, and therefore can't be extrapolated to predict near-term progress on Moravec-hard problems like robotics. What LLMs do is, in fact, Moravec-hard. And we should expect that if we've got enough compute to make major progress on one Moravec-hard problem, there's a good chance we're closing in on having enough to make major progress on others.

The US has a 'Do Not Call' registry for unsolicited phone calls, but technically doesn't need one for texts because it's illegal to send marketing texts without prior consent in the first place. Thing is, 'consent' often just means failing to notice a checkbox during a signup flow or something, so people end up getting junk anyway.

Even more annoyingly, politicians wrote in an exception for themselves. In combination with the way campaign finance works in the US, this means that if you've ever give your number to any political campaign, it will be passed around forever and you'll have multiple politicians begging you for money for months leading up to every election. Each individual campaign/organization seems to respect 'STOP,' but once your number is on an e.g. 'Has ever donated to a Democratic candidate' list, there's seemingly no way to get it off for good. Thanks, Obama. (I gave him $50 in 2008.)

The irregularities of many real-world problems will keep even humans of low intelligence employable in non-AGI scenarios. Consider that even if you build a robot to perform 99% of the job of, say, a janitor, there's still that last 1%. The robot is going to encounter things that it can't figure out, but any human with an IQ north of 70 can.

Now, initially this still looks like it's going to reduce demand for janitors by 99%. So it's still going to cause mass unemployment, right? Except, it's going to substantially reduce the cost of janitorial services, so more will be purchased. Not just janitorial services, of course. We'll deploy such robots to do many things at higher intensity than we do today, and as well as many things that we don't do at all right now because they're not cost effective. So in equilibrium (again, the transition may be messy), with 99% automation we end up with an economy 100x the size, and about the same number of humans employed.

I know this sounds crazy, but it's the historical norm. Today's industrialized economies already have hundreds of times the output of pre-industrial economies, and yet humans mostly remain employed. At no point did we find that we didn't want any more stuff, actually, and decide to start cashing out productivity increases as lower employment rather than more output.

An AGI can presumably control a robot at least as well as a human operator can. The hardware side of robotics is already good enough that we could leverage this to rapidly increase industrial output. Including, of course, producing more AGI-controlled robots. So it may well be the case that robot production, rather than chip production, becomes the bottleneck on output growth, but such growth will still be extremely fast and will still drive demand for far more computing capacity than we're producing today.

Demand for compute will skyrocket given AGI even if AGI turns out to be relatively compute-efficient. The ability to translate compute directly into humanlike intelligence simply makes compute much more valuable.

In his recent "Intelligence Age" post, Altman says superintelligence may be only a few thousand days out. This might, of course, be wrong, but skyrocketing demand for chips is a straightforward consequence of taking it seriously.

Non-general AI won't cause mass unemployment, for the same reason previous productivity-enhancing tech hasn't. So long as humans can create valuable output machines can't, the new, higher-output economy will figure out how to employ them. Some won't even have to switch jobs, because demand for what they provide will be higher as AI tools bring down production costs. This is plausible for SWEs. Other people will end up in jobs that come into existence as a result of new tech, or that presently seem too silly to pay many people for — this, too, is consistent with historical precedent. It can result in temporary dislocation if the transition is fast enough, but things sort themselves out.

It's really only AGI, by eclipsing human capabilities across all useful work, that breaks this dynamic and creates the prospect of permanent structural unemployment.

Self-driving vehicles can be symbiotic with mass transit. Autonomous taxis can make rail work better at low to moderate densities by shuttling people to stations. They can also ease the path to higher density development in car-centric suburbs by dropping people off at their destinations and then going to serve other riders or park out of the way, eliminating the need for density-killing parking lots immediately adjacent to businesses. As density rises in these areas, building mass transit will become more viable.

Nor does autonomy have to be limited to car-sized vehicles. It could be used for larger vehicles, perhaps intermediate between car and bus size — once you don't have to pay for a driver, operating a larger number of smaller vehicles is much more viable. These could be dynamically dispatched and routed, eliminating many drawbacks of existing bus service.

In a similar vein, LessWrong released an entire AI-generated album with lyrics adapted from significant posts made there over the years: https://www.lesswrong.com/posts/YMo5PuXnZDwRjhHhE/lesswrong-...

I think I'm going to enjoy how surreal widespread access to generative AI will make the world.

This is likely achievable in theory now, at least between major airports with ILS-equipped runways, but fully automated systems couldn't handle present traffic coordination procedures. You'd need a series of new standards to replace human-oriented air traffic control with a scheme in which ground computers could directly interface with aircraft flight management systems. Developing something like this and rolling it out on a global basis, in such a safety-critical application, would likely take two or three decades. Not clear it's really worth the trouble, since you'd want backup pilots for unexpected situations anyway.

What would probably make more sense is to just add a single-button auto-land feature, that sends an emergency destress call and configures and invokes existing automatic navigation, approach, and landing features to find the nearest appropriate airport and land. Given how rarely this would be used, there wouldn't be a need for the system to navigate complex traffic patterns, as ATC could just clear other aircraft out of the way. Something like this has recently become available for general aviation aircraft, but I haven't heard about anyone working on it for airliners yet.

"Code commands" is very plausibly an autocomplete flub of what was supposed to be "code commits." When I type "code comm" my iPhone offers up "commands" as the completion.

I've seen a lot of mockery of this request, but I suspect people aren't considering the wide variance in employee quality that can exist within a mismanaged organization. What Musk was asking for here wouldn't be a good way to evaluate skilled, conscientious developers, but it would be a pretty effective way to rapidly identify people who are basically incompetent or just aren't really doing anything.

The feature only works on P3 displays, so gamut when displaying UI (which I'd assume is nominally within the smaller sRGB space) likely isn't an issue.

Also, per later comments, this feature only kicks in when you're actually viewing SDR content, so there's no downside (even contrast) in everyday use.

On conventional LCD displays you are presumably losing some contrast in SDR content here, because running the backlight brighter to allow for this will also elevate black. The rumor mill suggests Apple is planning a move to Mini-LED displays with thousands of local dimming zones, which would handily address this.

Centralization significantly contributes to reducing friction and increasing user trust. Note that even on platforms that do allow competing app stores, market leaders can command similar percentages (e.g. Steam on Windows). This is why.

If Apple opened up the platform, developers wouldn't be paying 5% to Uncle Bob's Discount App Store instead of 30% to Apple, selling just as many units, and pocketing the difference as profit. They'd quickly discover that users preferred to buy from one or two trusted app stores, and that the better move was to pay whatever they had to pay to be present in those stores.

Are companies that provide one-to-one communications services, like cell phone companies, complicit in drug dealing, insider trading, and all the other nefarious acts that are no doubt organized using their networks? I think most people would say they're not. Meanwhile, one-to-many communications services, like broadcast TV, by their nature require editorial decisions. If you're only broadcasting one stream someone has to decide what the content is, and thereby becomes responsible for that content.

Social media is many-to-many. Quite a few people seem to have decided this imposes the same sort of centralized responsibility that exists with one-to-many communications, but this is not at all obvious to me. On social networks there is no single stream of information, thus there is no necessity for editorial decisions, thus the mechanism by which one-to-many broadcasters become responsible for the content they carry is not present.

Recommendation systems complicate this somewhat, admittedly. If platform owners are putting their fingers on the scale (algorithmically or otherwise) to determine what gets exposure, some of the responsibility comes back. However, in the simple case of e.g. Twitter showing you content from people you've explicitly chosen to follow, I don't see how they've got any more responsibility than exists in the one-to-one scenario.

Pretty much nobody prices software or digital services by looking at how much they cost to provide and adding some margin, which is how this implies Apple should be reasoning. The common model is to price on value. Taking a percentage of revenue is a pretty good proxy for that.

And Apple absolutely does increase your reach. Not in any important technical sense, but by creating a low-friction, high-trust environment, which results in a much larger market for apps than would exist if purchasing required users to type their credit card numbers into your web site and download and run an installer. This is a huge part of why native apps even emerged as a major phenomenon on mobile, against the tide of webification that's swept the desktop over the last 20 years.