rad-b
rad-b on Hacker News
In the I was referring to people reading my comment as “us”.
While it is nice to think of better futures for coming generations, I think we (reading this) also have to think about our more immediate work, as in next 2-5-10 years.
And like to anyone else who hints that “automation creates better jobs”, I have to add: but not for us. Maybe for the next or the generation after that.
Industrial and technological progress has birthed better conditions overall, but at a cost of locally worse ones. Mine workers that got displaced didn’t end up being machine operators - most of them landed on benefits or sporadic gig work. Only the next generation was able to be properly trained and educated to be able to benefit from the technological advancement en masse. And we might see that happening again.
Still haven’t found an ad not bested by NextDNS. Granted, not an ad-blocker extension but it “just works” and blocks ads, so good enough?
Yes, exactly! Maybe not sarcastic but definitely using jargon and being opaque like we all know it, it’s just hinting. Feels like talking to someone who’s constantly trying to prove to me they are smart(er).
I don’t think that would go in consumers’ favor. If anything, it’s more likely that dynamic pricing is not hurting the protected groups as their members on average have less disposable income and hence see lower (non-inflated) prices.
Can’t wait for the onslaught of opinion pieces on “Has dynamic pricing fixed wage inequality” or similar. Less discrimination if an airplane ticket always costs 300$ or 3% of your monthly salary, whichever is higher.
All this AI fluff, and nowhere on the page do you state what the company does. I feel like I have read a personal AI journal and not a company page.
I’m sorry to see so many comments bad mouthing the article when it seems like they didn’t read it. I thoroughly enjoyed it. The fable is convincing and lays down the hierarchical trust pretty clearly. It could benefit more from the political part of economy development, but even just as is, it’s a great primer on some of the more important concepts in modern banking.
Very valid point, shame it’s buried so deep in the comments’ tree.
“People could only see the listing and request to book the property during a 20-minute window”
This seems like a much smaller problem than the title would suggest. I agree Booking fraud is worth tightening but hardly anu automate system would flag a listing that was open for 20min only, as the calculated risk/impact to users would have been very small.
It seems to me you underestimate the challenges of getting the money from such crypto scheme. A murderer would have to prove to the pot owner that they are, in fact, the real murderer and not someone pretending to be one to get the money. However, there’d be another actor trying to prove who the murderer is - the police. So, you’d practically have to dox yourself in order to get the money, or be able to have such a strong privacy system that a state actor cannot break it and identify you from your communications with the pot owner (lol). If you have such a strong cyber security system, you’re probably better off just selling it for $$$$$$ than muurdering randos.
But these problems would likely be self-resolving in most cases as the crypto pot owner would simply rug-pull everyone.
Agreed, a dorm-like setup with amenities in the middle might work. It’s a whole different economic model and pitch to potential buyers/renters, but it could work.
The key problem with adapting offices into apartments is the lack of natural light in office floor plans. You have a glass skyscraper that has a huge floor surface, but no way to deliver light in the middle of the building. So even if converting it, you’re looking at making very big apartments since they need to hug at least one exterior wall to have any light, and then go toward the middle of the floor plan, whilst getting darker and darker. Essentially you’d be buying a 1-window or 2-window apartment with multiple bedrooms, which is hardly desirable. The problem is that we’ve built office spaces quite differently than residential high-rises, making a conversion into (desirable) apartments sometimes untenable.
Do you use summarize email function in gmail or outlook? Do you use rewrite paragraphs feature in word or google docs? Do you know anyone who wants these features?
That’s what I’m talking about, not AI use in general.
I’m not saying nobody wants to use AI, or AI coding and desktop apps. I’m using that daily as many others here are too.
The article refers to different AI features within other apps (e.g. search overviews in google, summarize email, summarize message) that bring no benefit to the user but only clutter.
I think everyone is fine with using their AI of choice to do their tasks of choice. It’s when I see a button that obstructs my workflow prompting me to do something I don’t need to with a shitty model that’s not good at it.
Such a cool guide to such a bizarre problem we live in…
Companies forcing features that nobody wants, but that are also expensive to operate and cost real money to run? Market can definitely stay irrational for long.
I see the regulators only regulated that first-party and third-party apps be treated equally, and didn’t specify how.
It’s a bummer (albeit an expected one) to see Apple reducing the burden for third-party publishers to collect personal data, rather than increasing the burden for themselves to collect it, thus overall raising the floor of user privacy. This is to say, when a regulator requests “everyone be treated equally to ensure competition” maybe they should also take into account how users are impacted and instead request something more nuanced like: if there are unequal standards, they must be equalized in a way that maximizes user benefit.
Great read and interesting analysis! I’m less charitable toward Anthropic supposedly not distilling ChatGPT for training purposes. Maybe not today, but during the GPT-4 era when Anthropic was the underdog - I can see it happen. Packaged along with some of Amodei’s clever jumping through hoops to prove how that is, in fact, virtuous.
Fun game! Enjoyed the panicked looking for the best available seat.
Not the OP, but I understand the sentiment they’re trying to convey. I will frequently work with AI to get done something that is not in my immediate area of expertise. And this kind of speak constantly happens, where you can feel it trying to please its RLHF expert teacher by being as concise as possible. Works ok when you’re deep in the weeds of what you’re talking about, but otherwise just feels overwhelming. Kind of like a feeling you can get when asking someone who is more interested in proving they are smart and know the answer than they are interested in actually helping you…
And yes you can prompt it to take it easier as you’re not experienced in this, but it’s usually a boiling frog situation. You work on something, need a small thing ajdecent to it, and three messages deep Claude is speaking to you in tongues unbeknownst. You knew what was happening at every turn but still got lost somehow.
How did you decide on Pi over Opencode? What you see as the “edge” Pi has?
It seems to be only against ToS; the problem, however, is that “distillation” involves intent. Anthropic seems to be calling for ban on distillation because they cannot reliably identify who does it - if they could, they would’ve just banned all infringing users.
However, if there is no systematic difference between regular use and “distillation use” then they need to establish intent, which is laughably hard if you cannot even distinguish use cases.
Hence the call to make it illegal too (and just against ToS). Because if it is illegal, then they can rely on US to enforce the law. And the US, for a lack of technical solution, can use the good old “China bad” shortcut.
Good idea, and unfortunately one of those that would work only while it’s not popular. If it turns out to be good enough to become popular, here comes the enforcement.
So refreshing to see such and interesting innovation in the hardware space! Curious as to what’s the business play here, as at $1400 a piece these don’t seem like mass-market devices. And the high-end market is even more constrained with the size of the subpopulation it is aimed at. Regardless of the business model, fingers crossed that more great assistive tech like this one ships!
Great comparison! We only have to take into account that applying lipstick well bears no consequences, but applying it poorly (i. e. new model tanking the benchmark) could amount to potentially losses of billions of dollars for the pig-breeders (AI labs).
A genuinely interesting read, thanks for sharing. Feels like GTM is always the hard part, regardless of how well you solve the technical piece. You can have the best solution in the world and no one to sell it to.
That might seem like obvious pivots in hindsight , but there’s no way to make such easy predictions going forward.
“Jensen’s job is to just be one step ahead of the market” — how is that a “just”? That’s arguably one of the most difficult things to do reliably (judging by how much payout it can give you). Feels like saying he’s making “just” billions of dollars.
Fully agree, that’s how I see most of my less technical coworkers reason about using AI. “Is there a Claude skill for that?” is a question I hear multiple times a week.
I see a lot of comments (incl 2 sibling comments) are always discussing whether there is a moat on AI. We agree there is no technical moat, there is nothing that Anthropic or any other AI lab could do that wouldn’t be quickly offered by other competitors too. However, _market penetration is the moat_. The deeper Anthropic is in relationships with orgs, the higher the cost of switching. Sure, for an individual it’s a 2-second job, but for a business it’s actual work of changing permissions, provisions, updating vendors etc. Nothing catastrophic but still real work, implying that Anthropic would need to drop the ball significantly to be swapped out, and wouldn’t be just because there’s a competitor who does everything kinda the same for kinda the same price (or slightly less).
Moat can be in execution and not just in technology. McDonalds has no specific burger-making technology that no other restaurant can acquire, what they do have is a well-scaled execution. Sure, there are competitors, and sure there are new comers to the burger space with different recipes (e.g. smash) but that doesn’t mean McDonalds is going under. Market penetration is the moat.
Unfortunately this seems quite plausible from today’s POV. As the old saying goes if you don’t want to be the product, you’ll have to pay for it. And I see only a silver of people being rich enough to afford and educated enough to care for paying privacy- or sanity-preserving tools and services.
Working in selection, I can say it’s more nuanced than that. Any measurement can be used as long as it is relevant to the business and related to performance. For example, you’re fine to reject people based on height if you’re hiring basketball players and being higher predicts scoring more points. Or even reject people based on gender (or other protected classes) if you can demonstrate that that specific group is absolutely necessary for you e.g. you want a counselor working with sexual trauma survivors and have evidence that matching patients to counselor on gender gives meaningfully better results for said patients.
The specific cases you mention and the finer point is how do you demonstrate the necessity of a measure? Is high general IQ absolutely necessary for SWEs? Or is it enough to have a high logical reasoning, but don’t need spatial? Do you really need high IQ or is it enough to have a lot of practical experience with hands on skills? Do you need higher IQ to do zero to one development vs code maintenance? The devil’s always in the details with these kinds of questions, and it’s definitely not a blanket “you can’t use anything”.