Hacker Newsnew | past | comments | ask | show | jobs | submit | jstanley's commentslogin

I noticed the same thing after I started dressing smarter.

If 0 and a function that always returns 0 are the same thing, does that make `lambda: lambda: 0` also the same? I suppose it must do, otherwise `0` and `lambda: 0` were not truly the same.

In a non-strict language without side-effects, having a function with no arguments does not make sense. Haskell doesn't even let you do that.

You can write a function that takes a single throw-away argument (eg 0 vs \ () -> 0) and, while the two have some slight differences at runtime, they're so close in practice that you almost never write functions taking a () argument in Haskell. (Which is very different from OCaml!)


Yes, and that becomes more intuitive when you "un-curry" the nested lambdas into a single lamba with twice the number of arguments. The point is that the state of a constant does not depend whatsoever on the state of the (rest of the) world, how much ever of that state piles on.

Another way to make the point: when you write 0, which do you mean?

In a pure language like Haskell, 0-ary functions <==> constants


I also understand that in the US it is the etiquette to cut your food up all at once, and then put the knife down, and then move your fork to your right hand, and then eat all the pieces using just the fork.

Use a knife and fork

As a paying customer, I don't care where the model comes from, I only care how good it is.

Sure, and also at what price point.

But can I rely on Cursor to be able to keep delivering, when they aren't the one's doing the work themselves?


Can you ensure that Notion is able to keep delivering given they don't develop their own models? Lovable? OpenCode? Should we be worried that Discord might disappear because they don't run their own data centers? Personally, I'm very concerned that one day Google might just have to close up shop, because while they do design their own chips, they don't fabricate them in-house; and don't get me started on TSMC and their critical dependency on ASML, they might as well just lock the doors.

Well, they can keep stealing as long as someone open weight their models.

And how cheap it is

> As far as I can tell, there is nothing about the training process of these models that would encourage them to make the output of any layer apart from (n-1) meaningful as the input of layer n

Right, I had the same thought.

Even if the output was in the same "format", does the LLM even have any way to know which order the outputs will go in? The ordering of the nodes is part of our representation of the network, it's not fundamental to it.

It would be like shuffling the bytes in a PNG file and expecting the program still to understand it as a PNG file.

The more I think about this, the more I don't get this at all.


These layers are residual layers, so what a layer does is:

x = x + layer(x)

so it's not too surprising that they can be used recurrently


Ah! Thank you

It's all new to me.

If you're time-poor maybe you're not as rich as you think.

The richest person I know talks to robots all the time.


> The richest person I know talks to robots all the time.

I've noticed this too, but I always thought of it as mostly people fooling themselves.

If you're rich (let's say anywhere above 10mil), it's practically guaranteed that you can allocate resources in such a way that more effective engineering, or science, or whatever, is done in less time than if you tried to do it yourself (rather than spending your time allocating capital). I've actually thought of this as a bit of a curse: the value of a rich person's labor output is inverse to their net worth. No matter how smart, you're not smarter than a crack team of Ukrainian/vietnamese/taiwanese/Indian scientists/engineers/whatever, and the more rich you get the more you can stack your crack teams, either paying higher salaries for higher skilled people or building bigger teams.

I think there's maybe 100 outliers to this rule in the world, people like John Carmack. I mean I assume he's rich.


I don’t think John Carmack likes to tell people what to do, regardless of wealth.

https://news.ycombinator.com/item?id=26170052


I'm not sure that he doesn't like to, so much as that the position he ended up in as a result of the Oculus acquisition had no actual authority attached to it. He was functionally a glorifier adviser, to trot out at trade shows (and reading between the lines, this was a pretty frustrating position to end up in - he'd rather have had a real job, even if it was to build something he didn't fully agree with)

So what, the richest person I know talks to DMT jesters, it doesn't make it good.

The richest people I know talk to a range of people like personal assistants, but really the PA is valued for getting things done reliably and in the real world with any needed resources. Even calling in experts as needed - of course they may indeed talk to an AI too

Nah, they're right. In fact, "self-service" is one of the biggest value transfers from people to capital owners, a society-wide "fast one" the computing industry pulled over everyone.

It's cool that you can do something yourself with a computer, whether it's ordering food or picking clothes or booking a trip. But, market doing market things, that can quickly became a have to, which is much less cool.

It's a problem that's hard to see until you're certain age (and therefore easily dismissed as whining of old people yelling at cloud(s)) - it's because most people in the west start with no money and lots of free time to burn, and gradually become extremely time-poor as their start working and accrue responsibilities (and $deity forbid, start a family).


Correct.

The smartest people in academia get promoted to positions that used to come with administrative staff.

Now they’re expected to do all of that with a computer, which is easy right?

So now they spend 30% or more of their time administrationating their position, rather than delegating those duties to their admin staff.

That’s less time teaching and innovating.

Meanwhile, the increase in administration costs of learning institutions has massively outpaced all other costs as a fraction of total.


Same is true in all white-collar work, too. I mean, not to look too far, it's very obvious in our own industry if you look for it. Highly-paid engineers hired for high-skill engineering work, but spending most of their time doing their own task management, calendar management, memo writing, presentations, trip planning, trip expensing, filing HR documents, and such? Heck, even the proliferation of ideas like "devops" or "devsecops" or whatever-ops, lauded as breaking down siloses, is just using buzzwords as cover for another iteration of headcount reduction.

My company won’t backfill product management in timely fashion.

Guys this isn’t an optional position. You don’t want your SWEs doing product work. They are not going to do a good job of it when they also need to, you know, do their actual job.


Yeah, the bit where there are 10x as many administrators in higher education, but professors now all have to do their own admin, always drove me up the wall

That is more money for the baseball team

All of the demos of booking travel using AI are hilarious to me. This used to be a job a travel agent did, and planning a trip was either a fun conversation or you could be like "send me somewhere warm" and let them do it.

Is it cheaper now that you can swear at flight booking software yourself, and scream at the hotel when they cancel your rooms that you got from a third party site that went through some other intermediary that bought the rooms at a group rate they shouldn't have been allowed to buy it at? Sure, it's cheaper. Is it better? Well, they want you to believe that. You have unlimited choice now. Oh sure, all the web searches and ads are targeted in a way that you're going to end up at the same place a travel agent would have put you, but you can perceive the freedom of choice along the way!


> Oh sure, all the web searches and ads are targeted in a way that you're going to end up at the same place a travel agent would have put you, but you can perceive the freedom of choice along the way!

And you can enjoy all the risk and liability for mistakes made along the way, too, which is where the actual optimization happened in the economy.


Similar example in the grocery stores with the self checkout. In the past if the employee did a scanning mistake, worst case the manager / customer would be mad.

Now that you do it yourself if you mis scan organic tomatoes as regular tomatoes you are freaking going to jail.

Ok exaggerating a bit, but having shoplifting in your record can be life changing, specially for immigrants


Well, even without exaggeration - if the employee made a scanning mistake, most of the time they (or the customer) would notice during or immediately afterwards, so the employee would just hit undo or scan a negative or such, and carry on.

No such privilege is granted to regular customers. Instead, the self-checkout station locks itself up, and the customer has to wait several minutes for the assigned employee (who, most of the time, is also working two other tasks at the store) to show up, analyze the situation, enter service mode, and do the undo steps.


Do they ever actually analyze the situation? In my experience they just ignore any issues and hit "approve" and on you go. I could have done that myself.

It's a classic false-positive problem. Most times when the self-checkout clerk has to give you attention, the problem is stupidly innocuous, so they blindly approve, as they have been trained by the system that it isn't a real problem.

I'm sure plenty of things get by them this way.


Some times I'm curious to see how stores work at the US nowadays.

My experience is that the assigned employee is always looking for something to do, because he can't leave the self-checkout area, but there isn't anything to actually do there. And well, the store better not accuse honest customers of anything, or else some stuff they really won't like will happen (and that applies to poor customers too).

Anyway, the experience is still so bad that I tend not to use it. But that's because the machines really suck.


Between Costco/Target/Winco/Walmart/Home Depot/Lowes/Kroger/Uniqlo, my experience is that I can check out quicker than before. I rarely have to wait for assistance, which itself is rarely needed.

I greatly prefer the single queue in self checkouts rather than betting on which cash register line will get stuck on someone that has a pricing issue or something. Obviously, this has nothing to do with self checkouts, but I find single queues far more ubiquitous after self checkouts came around than before.

For lots of stuff, a cashier is probably quicker. But I almost never have lots of stuff.


> Now that you do it yourself if you mis scan organic tomatoes as regular tomatoes you are freaking going to jail.

If this happens, it’s a problem with the judicial system, not self checkouts. I highly doubt it has ever happened, though.


Just buy insurance! Oh, it's up to you to understand what it actually covers, and it's about as much as the room/flight costs but won't you feel better about your choice?

> And you can enjoy all the risk and liability for mistakes made along the way, too, which is where the actual optimization happened in the economy.

And you can enjoy overpaying for a lower quality vacation if your travel agent is unscrupulous and getting kickbacks from vendors.

That is the actual optimization that happened, I can do more research and communicate directly with vendors.

Those that don’t want to are still free to pay extra for a travel agent.


> That is the actual optimization that happened, I can do more research and communicate directly with vendors.

Only if your time is worthless.


What did travel agents do when they made a mistake? I don’t think they reimbursed people?

There's multiple levels missing when you do it yourself. Usually they would do their best to sort things out. At the very least it was a single call to deal with the issue to someone who knew how to rebook flights, find new hotels, whatever, not you struggling to figure things out on your phone. For a full mistake, yes, they'd usually reimburse you, if not there are regulations around refunds and things like small claims court.

Doing it yourself? Good luck! Hope you've got good service on your phone where ever you happen to be when things go wrong.


>Doing it yourself? Good luck! Hope you've got good service on your phone where ever you happen to be when things go wrong.

90% of travel is probably happening where mobile networks are available. Also, since most travel seems to happen without travel agents today, it appears that "luck" is not that necessary, otherwise people wouldn't be choosing to forego travel agents.


They were in a position to notice and correct most mistakes near-immediately, or at least shortly after making it. For most other cases, apologies and/or reimbursements backed by insurance if needed, transparent to the customer. In self-service, all that is responsibility of the user, but it's all built on requests to third parties, so the user is not in a position to unilaterally fix a bad request.

The travel agent is also not in a position to "unilaterally" fix a bad request, they are also requesting other parties to do things.

Travel agents were not outlawed. Most people just prefer to save money and do the work themselves (for most trips) rather than pay a travel agent.


> The travel agent is also not in a position to "unilaterally" fix a bad request, they are also requesting other parties to do things.

Yes, but they're already one level up, so they can fix the problems in their company's immediate system, and then unlike the customer, they're a trusted party in the network of all other parties, so they can mail/call other parties directly and get people there to fix issues without too much delay.


By choice. Your friend is presumably wealthy enough that they could talk to a human instead, or completely delegate whatever they’re talking to AI about and never talk of it further.

Oh I am not speaking from experience here, I'll clear that up.

Also the original saying used rich people but I think it better pertains to busy people in general.


Is this down now or am I just too stupid to work out how to use it? It isn't generating any translations.

Firefox on Ubuntu.


Did you touch the Cloudflare captcha in the bottom left? Once you pass it you need to trigger another change event (just type a space)

> What are we trying to achieve as a species?

How often can a group of even 3 people come up with one thing that they're trying to achieve? You're not going to get consensus on what the goal is for humanity.

> If we view humanity as the most intelligent life form not just on the planet, but in the entire universe, then we are the only source of order in an infinite chaos.

That doesn't follow at all!

And even if it did, why stop at the species level? If humanity being the most intelligent species makes humanity the only source of order, does Terence Tao being the most intelligent human make Terence Tao the only source of order?


> why stop at the species level?

The stop isn't made at the species level, but at the intelligence gap. The gap between Tao and a random uneducated, malnourished human is nothing compared to that between the latter and any other animal known to science, or of course to the stones and flame occupying the rest of space.


Well, any group of 3 people would easily agree on one thing: that they are all just trying to survive. Now, if only what we’re doing to nature is actually killing us…

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: