>Also, I’m really unimpressed by any “one shot” demo that’s out there in the wild. It means nothing for serious software engineering.
If a person, or team of people, can build a demo quickly then it's good odds that they can build the real version (though, famously, not a guarantee). However, it turns out that a machine that can spit out 100 demos of whatever can't actually build the real thing.
Similarly, a chess engine rated to 1000 Elo doesn't play like a 1000 rated human being. The mistakes that each make to reach the equivalent level are different in size, frequency and kind. The thing that makes a human reach a good demo is very close to the skillset to reach the finished article. This isn't so for LLMs but we have yet to update our priors.
This resonated with me. "Developing ideas and artifacts using AI breaks our normal intuitions along many meaningful axes and we've yet to update" is a really clean idea.
One of my favorite delusions is that men are only better at sports than women because they are socialized to be so from a young age - that if women & girls were encouraged to be strong, fast and competitive; and if boys were scolded or perhaps asked nicely to play... board games or roleplay or whatever, then there'd be an inverse. Oh, and this delusion is comorbid with the idea that the reason sports are segregated by sex is so that men aren't embarrassed when they are beaten by women.
What's fun about both of these delusions is that they are easily disproven by even mild contact with reality. Any level of sports training with a member of the opposite sex disproves this with incredible speed. You can read social media comments from women about the first time they wrestled or playfought their mostly inactive boyfriend for the first time, and asked them not to hold back. And this doesn't even need them to engage in any kind of sport.
To hold this, you have to wilfully, intentionally dodge all physical activities with the opposite sex from schooling to your adult life. And yet, it does exist! I've met people who believe this, in the face of all reality. And then these sports-dodgers will have consequently wrong opinions on e.g. transwomen in sports, whether transmen should be allowed to take steroids and compete, equal pay in sports...
That said though, I could believe there's some crossover point in ultra-ultra endurance sports where performance equals out between the sexes.
i recall mcenroe said serena couldnt be considered greatest of all time in tennis regardless of sex because she wouldnt even beat a man ranked inside the top 100. he said she'd herself be ranked somewhere at 700.
we arent even talking about women only asked to play 3 sets vs men playing 5 sets (at least in slams).
I think McEnroe was very unfair, and that men and women can be completely equal in technique and skill, even if not in physicality. But in my opinion, the women's game is what you watch if you want _even more_ tennis, or perhaps you really like the story lines of your favorite athletes for whatever reason. Similar to watching Formula 1 is just being a regular sports fan, but watching the Formula 2 season marks you as a slightly obsessive motorhead.
Semianalysis is an AI hype organisation, not a serious, unbiased semiconductor reviewer/journalist like chipsandcheese nor a documentarian of the semiconductor industry like Asianonmetry (as it relates so strongly to the modern economies of Asia). If you see something from semianalysis, you can simply ignore it.
Semianalysis seem to have useful information but increasingly crazy extrapolations of trends and future predictions.
It's helpful to realize that Dylan (Semianalysis), Dwarkesh, Aschebrenner (the Situational Awareness guy) and Sholto Douglas (Anthropic) all share a house in SF, so what you are getting from any of them is the SF AI scene view of the world, which is interesting to know, but probably not the best predictor of how things are going to pan out.
I tend to agree, but the Semianalysis + Dwarkesh side of reporting still does surface interesting information. You just have to take it all with a grain of salt, as indeed it is more hype focused, and look for real information hidden in the noise. And possibly to be a bit more entertained as you do so.
I suppose access journalism does have to get something out of the bargain, however small it's not zero. If that's the way you like to spend idle time, well it takes all kinds. I don't understand competitive scrabble either.
The trouble is the article is so poorly written I don't want to look for hidden information, i want to close out. I'm guessing ai wrote it, the information is constantly repeating, and not even always consistently
Yeah. I can't do the physics myself, so I take on faith (heh) that a sufficiently fast rotating blackhole has a ring shaped singularity that can extend outside its' event horizon.
This was news in the 90s and it was a plot point in at least two scifi books, though I don't think I can recommend either
>Why do these otherwise very smart people always seem to draw these kind of shortsighted conclusions?
The obvious conclusion is they aren't very smart. Consider that Cluely guy: the best returns on lying are to be the only one lying while everyone else tells the truth, so why does he want to promote lying? Partly dumb ragebait marketing (this isn't going to attract talent... and while it has given them millions in startup funding, how long are the sharks giving him the money going to do that?) and partly he just isn't very smart.
You gotta remember that people with money have money. Money doesn’t make you smart, gives you the ability to hire the smartest people.
But, to paraphrase the Prussian general 2 x 2, the most dangerous person in power is a man who is both stupid and diligent. When people spend a lot of time smelling their own farts, they start believing that it’s perfume.
Amazon is almost completely unusable for search at this stage. I think perhaps 3/4 of the results are sponsored ads. Three in four. Trying to find a good deal on a product, even if you know precisely what you're looking for, is a real headache of trying to dodge the advertisement minefield.
Even without the ads, Amazon search is abysmal. It seems to ignore half the words you enter. I often resort to using Google or ChatGPT to find items on Amazon which I know are there, but which their search box doesn't surface.
I searched for "cayenne" and accidentally bought the hot pepper product that came up which was not cayenne.
After going mildly buckwild on amazon after 3 years of not ordering a single thing (after a move to the country) I let the free month of prime I was using expire. This, and the counterfeit product I got (and had to spend 45 minutes on the internet and phone to get refunded) was decisive. I'll try again in 5 years.
> and had to spend 45 minutes on the internet and phone to get refunded
anecdata and all, but i've had the exact opposite experience - I've never spent more than a couple of minutes getting a refund, and it's always been less time and effort than returning something to a B&M store.
I imagine most readers haven't tried tiktok shop (or similar Chinese consumer shopping platforms).
ByteDance has reimplemented their now infamous video suggestion algorithm for shopping and it is _very_ effective.
Once the infrastructure and logistics of the Chinese eCommerce platforms improve in the US, it will be so much better that I imagine that Bezos only recourse will be to have congress ban them (a la the Chinese EV ban).
He will likely be successful in doing this and allow the US to slide further into modern technological irrelevancy at his own benefit.
I have used this shop, but I do not agree that the experience is a good one. It is abysmal -- the standard for tiktok shop / aliexpress / temu is to offer you all sorts of "deals" that never actually apply. "Oh, you picked this item? Pick another four items and all of them are FREE!!! FREE!!! Also you have to buy a fifth one. Oh, and this only activates once you've spent more that $100. FREE!! FREE!!!"
Way worse than Amazon because at least Amazon is just showing you products and telling you how much they cost.
Its hard to pin down specifically. As a general rule: When I land on the homepage of the TikTok shop, without touching my screen I _generally_ see something that I am interested in buying. Its like it predicts what I am there to search for, and its quite accurate.
On the Amazon homepage that has never happened. It usually just shows some banner ad for prime day, new amazon tv show/movie, then maybe 4 little thumbnails of the last product I searched for or whatever. I have to initiate the shopping experience, I have to pique my own purchasing interest.
I’ve bought some stuff off there and can tease out some aspects, though they may be after the fact rationalizations of subconscious effects.
The first is price. The stuff is cheap, as in so cheap I know it will be terrible and I still sort of don’t care. It’s $2, if I only use it twice I got my moneys worth.
That makes it super easy to hit Buy without thinking much.
The second is that they appear to target verticals that don’t have a natural ceiling to sales. I only need one microwave, so they won’t sell me that. They do sell a lot of apparel, decorations, toys for adults (crappy 3D displays, not dildos), stuff where you don’t think “eh, I already have one though”.
The third is the marriage of entertainment and shopping. I can be doomscrolling, see an ad and buy it without ever leaving the platform. If I remember right, it doesn’t even interrupt your doomscrolling really. Swipe to next video, see it’s an ad, hit buy, confirm Apple Pay pop up, scroll to next video. You can decide you want something, buy it, and be on your next video in 15 seconds. It’s so fast it almost blurs the line where I’m not sure if I’d call it shopping or entertainment.
The last is that they’re really good at luring consumers to start selling on there. The ads for stuff tend to be sort of home made and low quality, which makes it seem attainable for your average person. That tends to mean a huge selection. I had a buddy that randomly decided to sell stickers on there just based on watching other people do it (no idea how that went, I didn’t ask)
> The third is the marriage of entertainment and shopping. I can be doomscrolling, see an ad and buy it without ever leaving the platform. If I remember right, it doesn’t even interrupt your doomscrolling really.
No trolling: This prose is wonderful. It reads like something out of a sequel to the film "A Scanner Darkly".
I hate videos when shopping. Videos are not skimmable. I don’t want to watch and listen to a salesperson sell the product to me. I want at least a modicum of individual agency and choice. I already skip all product videos on Amazon. I hate the product images too as they have become too salesy and they sometimes no longer even show important facts printed on the packaging itself (like the ingredient list or Nutrition Facts / Supplement Facts).
Fair; While I can only speak to TikTok shop (as its the only one I've interacted with), the videos typically take the place where product photos would be on the product page. A swipe dismisses it and its replaced with the product photos. The core browsing experience does not show videos in lieu of product photos from what I have seen.
Yeah, it's truly awful if you're looking for something specific. I tend to find things by finding something I don't want then clicking through 16 pages of "related to this item" ad nauseum.
I find myself more often using ChatGPT to find things now because their search is sooo bad.
Like, if I want petite clothes, I do not want clothes that are not petite, folks. Why is this hard?
I’m surprised that anyone would love Amazon so much that they resort to using third-party search so they can continue buying from Amazon. Rather than just shopping elsewhere!
There are many product categories where I know that shopping on Amazon will be a miserable experience riddled with crap goods and useless search, so I go elsewhere. Never even occurred to me to use an LLM to make up for Amazon’s own deliberate decision to serve me crap.
Eh, for me it's not about love, it's about no other retailers carrying what I want, at a price point I'll pay, with an easy return policy and easy checkout. I'm willing to put up with a lot of crap to find that, apparently.
To be clear, when I do use LLM for search, I don't limit it to one site.
It’s wild. I was trying to find a cheap bookshelf with a specific style the other day. Firstly, I couldn’t filter by exact width. It autogenerated a width range (like 24-30”) which was different on subsequent searches and several inches away from what I wanted.
Second, I found a product I liked but was the wrong size. As far as I could tell, they can’t even show you recommended similar products. They show you what other people clicked on or bought in the same order. But not just actually similar products in the same style.
I looked at well over 1000 listings when looking for a new desk. The filters are sick joke. This was pre-llm, today I'd write something to scrape the results and parse the product pages/images to filter out the crap.
This plus the login-walled reviews is unfortunate because once I stopped buying from Amazon I'd continue to go there to browse before buying elsewhere.
They also keep making it more annoying to even get to and read the reviews. It feels like it gets more hidden at the bottom of the page every few months. What a shame! Imagine how useful it could be.
Also, even after you find precisely what you're looking for, there is no guarantee that what arrives will not be a cheap imitation or reduced-quality version of what you purchased.
Well at the moment AI hasn't made anyone any money. It's redistributed investors money (typically wealthy although some pension plans have been caught up too) to tradespeople, to Nvidia, SK Hynix, Micron and other specialised vendors with pricing power.
After that it's like... Who is making money? "we're doing 100x the things we're doing before" people can't really point to anything. Maybe it's in use everywhere all the time, and they've been pumping out twice the amount of features that no one uses or asked for as before, in which case this is classic value destruction
Yes, they're redistributing my 401k and a good portion of the entire economy into it, with the ostensible goal of automating my job and putting me permanently out of work.
we are seeing robust profits and growth in margins in the S&P even ex tech and some of this spillover from the hyperscaling but some may be due to productivity enhancement from AI.
We have a lot of competing frontier or near-frontier AI companies. So far a lot of the productivity improvements seem to be delivered as customer surplus.
Worker pessimism over technological unemployment risk is enough to explain profit growth via wage suppression with no need to invoke productivity gains, and AIUI these sorts of indicators have been leading on AI capabilities and adoption.
There is money to be made in AI, the loss we're seeing is because they're in a bubble of spending way too much on R&D and building datacenters. AI companies aren't losing money on individual queries.
I agree the projected revenue makes no sense, but if the bubble bursts, we'll still be left with a technology that (1) many people want to use (2) is lucrative to the providers.
If a person, or team of people, can build a demo quickly then it's good odds that they can build the real version (though, famously, not a guarantee). However, it turns out that a machine that can spit out 100 demos of whatever can't actually build the real thing.
Similarly, a chess engine rated to 1000 Elo doesn't play like a 1000 rated human being. The mistakes that each make to reach the equivalent level are different in size, frequency and kind. The thing that makes a human reach a good demo is very close to the skillset to reach the finished article. This isn't so for LLMs but we have yet to update our priors.
reply