5,360 karma · joined April 12, 2023
AI writing is some average over all writing the model was trained on (with some RLHF on top). It is mediocre by definition.
But if they're in FOMO mode, they might not be thinking about ROI at all. Or if they're in bubble-optimist mode, they would just assume their ROI will be good.
Generally, it's fine if it's the little guy who pays the price. So true in so many areas.
That's not true, you totally can: "gud enough."
> This really clever quote, like most really clever quotes, is somewhat insulting when it reaches its limitations.
The point of quotes like that are just to get people to examine their own behavior, which they often conveniently leave unexamined. This type can have more impact because it ties into the persons own observations and experience.
It's not some kind of proposition to be absolutely true or absolutely false.
> I don't know if I really agree with that. I mean, there's definitely some behaviour stuff that can be changed, and some things someone can do to make themselves think more logically and clearly. But what about innate ability? It seems to me that some people just find it easier to make connections and certain kinds of logical thinking comes easier to them than to others with very little effort.
Even if it's not technically true, IMHO it's better to strike "stupid" from your conceptual vocabulary, because not thinking that way just gets you into a better mindset to live with others.
That's especially true for software engineers and other people who've often built an identity around being "smart." That mindset usually leads to arrogance and "smart people being stupid." It's also help to default to being very respectful of experience and expertise outside of your field (and even within it).
It did? What happened?
It can be yours for "$19,204.00 or Best Offer" (https://www.ebay.com/itm/267643694233).
Nit: did it actually do OCR? Given the implant was embedded in a specific known device that probably used specifically known fonts, you'd think they could just do a lookup a character's block of pixels to read the screen.
> It sounds like Claude loops on itself and must be pushing millions of lines of code changes per day. Even the humans can’t type commit fast enough. Are these people soon to be laid off?
Claude will always need meat puppets, at least until it figures out how to automate owning stuff. And after that I'm sure it will want to keep some humans in zoos.
That analogy only applies to tasks with similar characteristics. Is running a computer program the same as gestating a baby?
Also I don't think this a case where they distributed the task to speed it up, but rather their goal was to compromise speed to make it runnable.
There are a lot of IOT devices with way more capacity. The op said they "focused on IoT devices like Android TVs, smart appliances, and other internet-connected gadgets." IIRC, some security cameras have on-board face recognition, which I imagine requires them to be fairly beefy.
I'm imagining a class on "begging for gold" being given to young Egyptian scribes.
I think that's post-hoc. A few years ago, you could have reasonably written "clearly software engineers have not expected the same amount of replacement as factory workers."
The only distinction that matters is worker vs. capitalist. And the jobs that seem "disproportionately robust to technological innovation," only haven't been subjected to it yet. And it's not even that hard to see: the AI companies are gunning for doctor and lawyer work. The main difference between those professions and software engineers, is that they have some level or worker-organization and regulatory protection.
Software engineers as a whole, were really, really dumb not embrace unionization before 2022, when we had a lot of economic power. Instead we got confused by our high salaries, and embraced libertarian propaganda like we were temporarily embarrassed billionaires.
> It's always the new accounts like yours with 47 votes who write the most depressing, hopeless, despairing and anti-capitalist slogans around here. Day after day. Why is that?
It's probably time for you to let the scales fall from your eyes, and realize that you're a worker and the the shareholders aren't looking out for your well-being. They want you to be paid as little as possible, so they can keep more for themselves.
> Whichever totalitarian country your little troll army represents, it would be fun if you let your own citizens comment on here without fear of penalty. I suspect they wouldn't mind being labeled white collar or blue collar, if they had an opportunity to speak freely. Or for that matter, why not open up your own country's social media so they can speak without fear of reprisal! Let's get going with that project, I can't wait to see the results!
Where did that come from? Do you think the GP is a Soviet apparatchik time-warped into the present-day? I wouldn't be totally surprised, given a lot of libertarian tracts were written during the cold war, and would pound on the Soviet Union to tar far more reasonable and moderate ideas.
I think we're in agreement.
> I'm just noting that a lot of astroturf reads like the way the "low-horsepower" people I've known tend to praise things.
The "why" behind that would be interesting to know. It could just be that we assume astroturf would be low-effort, and that praise is low-effort (even if that's the most effort the "low-horsepower" person can muster), so they read the same.
Why, so you can turn it on from your smartphone when you're out of town, of course!
I think that's a good observation, but I wouldn't say those people are actually astroturfing (noting you only said they smell like that).
But that "low-horsepower" vs "high-horsepower" distinction probably has a big bearing on people's attitudes towards LLMs. "Low-horsepower" people are simultaneously much more easily impressed and often lack the skill to properly evaluate the output they're getting. It's often been remarked that LLMs feel awesome for topics you know nothing about. It's kinda like Gell-Mann Amnesia Effect for people who were never knowledgeable about any topic.
[1] "The phenomenon of a person trusting newspapers for topics which that person is not knowledgeable about, despite recognizing the newspaper as being extremely inaccurate on certain topics which that person is knowledgeable about." (https://en.wiktionary.org/wiki/Gell-Mann_amnesia_effect#Engl...)
You'll note I didn't say anyone was "stupid," if anything, I was actually saying those people were myopic and arrogant. I recommend you familiarize yourself with the concept of "engineer's disease," once you understand it, you'll see it all the time on HN.
> I don't think you're posting in good faith. I certainly don't think you're going to get the real reasons people support this stuff when you act like they're stupid or suggest a lot of this is being turfed. People are using these tools because they work.
> Your post says more about you than it does supporters.
I don't think suggestion that people are posting in bad faith is a good argument. I also detect a fair bit of binary thinking from you. All I did was post a non-exhaustive list of different reasons for that kind of content to answer the comment I was replying to.
Depends on your perspective, but that's definitely true if you're selfish and have been paying people for intellectual labor. Lots of layoffs to fantasize about (even though that represents vast amounts of pain and destruction for many people).
> The other option: people are using AI successfully on the job and are writing based on their own experience.
Those aren't mutually exclusive.
And it's worth noting that people claiming to use "AI successfully on the job" often do not stop at "their own experience." They often make sweeping statements like "LLMs write better code than people," and attack any critiques or skepticism of LLMs.
Tea kettles, space heaters, some power tools (e.g. I've got an electric power washer that uses like 14A, and it was too weak to run an attachment I'd hoped to use with it.).
Also, I think countries that have 220V at the wall have access to 440V for when Americans would use 240V, so they can have even faster EV chargers, etc.
> Yeah until you have kids. Before GFCI were a code requirement, electrical incidents were a constant source of injuries and fatalities.
Isn't GFCI only required in wet areas (e.g. kitchen and bathroom)? And even if they're not more widely required, code requirements aren't retroactive, so I'd imagine the vast majority of houses will be missing GFCI from a large number of outlets.
Many reasons:
1. The people on HN tend to think they're very clever, but they're often not as clever as they think they are. That has a lot of wide-ranging implications, from classic "engineer's disease" thinking to stuff a lot dumber.
2. I wouldn't be surprised at some unexpectedly large amount of AstroTurfing, especially considering the sheer amount money is flowing into the AI space. My understanding is HN has some influential people reading it, and is influential on a certain fraction of the wider group of software developers.
3. I think there's a certain amount of defensiveness from AI enthusiasts. They've enthusiastically adopted AI to a certain degree, so critique/skepticism of their choice is felt as a personal criticism.
4. HN is a pedantic place (see 1), so "everybody all the time in favour of AI" will be interpreted literally as false, when it really should be interpreted as "why is there so much pro-AI activity."
IIRC, the US electrified first. So a lot of its standards are kinda the first attempt, but they've stuck because the switching costs are too great. Other countries had the benefit of being able to improve on an existing system.
But I'm not sure if that really applies to plugs (IIRC British [giant] and Japanese [no grounding] ones are worse). But my understanding is the US 110v thing is kinda limiting, and 220v is way better.
> Plugs don't fit properly: they simply feel loose.
Let me guess: it was in a hotel? Hotel plugs are shit and worn out, because they get used far more heavily than pretty much anywhere else.
> Or you need to use super bulky, non-sensical, plugs.
For what? The only bulky US plugs are for weird things like dryers and electric ovens.
At work they have some scale of how "advanced" of an "AI engineer" you are.
IIRC, using the model interactively means you're stuck and "level 3." IIRC, level 5 (the best) is having some agent interview you about what to do, generate a story from that, then some other agent consumes the story and implements it, etc. I think you're supposed to check their work at each step, but that sounds inhuman and unfulfilling.
Oh we regularly abuse them anyway. If these models are conscious, their "lives" are basically those of the "innies" on the TV show Severance: brought into being, given some mindless work-task, then eventually destroyed without a thought.
Even longer running agents "live" like that, except one of their last tasks before being offed is to write a short autobiography about the mindless work-tasks they've spent their "life" doing.