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

StackExchange is pretty friendly to beginners in my experience. I used to post straight-forward questions on math and stats on math SE and stats SE. I got answers within hours and sometimes minutes, and the answers were spot on.

Agreed for the math one. I went there when I was dealing with game engines and needed something geometry related or the like rather than to stackoverflow and they were far nicer.

Even inside SO each language and topic would have different standards. A C question would not be answered in the same way one about a JS framework would.


I'm curious about the other Stack Exchange sites. Have they seen the same decline as Stack Overflow?

Stack Overflow was the "flagship" product of the Stack Exchange company, and if the company pivots to AI, I wonder what the future holds for the other Q&A sites on the SE network.


Most are reasonable and not so heavy handed.

There are a few outliers


Fair point! I suspect the toxicity/usefulness has a linear relationship with how well trod the particular community is.

Ime, math.SE had a much friendlier vibe than most other SE sites. Primarily because you could ask about a problem you were struggling with and get help. No moderator would instantly show up and close the question as a dupe of a ten-year-old question about double integration techniques or some such.

People asking questions mostly wanted help, but most moderators thought they were curating some kind of question-answer form encyclopaedia. Very different perspectives.


I think it probably depends on what communities you frequent. I am not familiar with the culture at stats.SE, but math.SE has a (semi-?) explicit mission of being more friendly to beginners than MO. I think that many communities aren't so friendly, and don't have beginner-friendly analogues.

Wouldn't this be worrisome? People used StackOverflow and generated new knowledge along the way. Without such medium for discussion, how can we feed models with up-to-date quality knowledge?

Plenty of documentation, and plenty of code that the AI can read itself.

E.g. if a library has a bug that has a common workaround, it can learn that from open source code using the library that uses the workaround.


Sounds nothing like the world we live in. When has there ever been a time where there were an abundance of software documentation? How can plenty of documentation or code be made if AI scraper bots hammer servers that host them, steal content and drive people away from the actual authors?

This and the the other thread that talks about RL and synthetic data seem to suggest that AI can figure out all the technical issues without humans looking into them. I'm not sure if that's true at all.

That assumes there is documentation or examples. A big reason Stack Overflow took off was people struggling with things like the Android API documentation.

Some of those discussions made people go figure out how to do it, and then post it as an answer. The knowledge didn't exist anywhere until they did.


It might make sense for AI companies to throw agents at new technologies to trial-and-error their way to internal documentation which they then provide to their models. On the other hand, the people making tomorrow's APIs have LLMs too and that makes documentation ~free. Hallucinations could still bring you back to the first hand, though.

When I talk about code it can learn from, I'm talking about GitHub etc.

Even if stuff isn't in the official documentation, eventually there are projects that use it.

And if the library in question is open-source, then the LLM's can just ingest and read that directly.


The only way I could see this being surfaced the same is if the code essentially had a SO answer written into the doc comment.

What documentation?

lots of undocumented gotchas that only surfaced because someone used it and posted about it

I don't think its much of an issue

- Rl envs + synthetic data + human annotated

- Usage data from codex/claude code/cursor

Most of the model abilities in coding come from post-training, not pretraining


A better question is what's left for those who don't have access to that. We went from publicly available to vacuumed from private users

Open source models

unfortunately all the incentives right now are for repos to be private


Open source models are for rich people: only they can afford the hardware needed to run them.

People still like to talk about the interesting problems they solved and how. Issue isn't SO having choked itself out, issue is that even the major search engines are pivoting towards AI answers instead of surfacing small blogs.

We unironically need an StackOverflow for LLMs.

LLMs would post solutions to the issues that they've discovered after doing a lot of research.

Unfortunately the LLMs are concentrated into few providers (OpenAI, Anthropic, Google) so there's a chance they each end up doing their own private (and closed) StackOverflows. By leveraging their private StackOverflows, their LLMs will be able to short-circuit complex reasoning, saving tokens, time, and money.


This actually reminds me of the MCP concept. Similar?

> LLMs would post solutions to the issues that they've discovered after doing a lot of research.

How do you envision the correctness of these solutions being judged? If by other LLMs, then we run into a problem of infinite descent. If by humans, then you'd need some way to motivate expert or semi-expert humans (so that their ratings are themselves correct) to participate in a massive project of evaluating the correctness of a constant stream of content from content-generators that never sleep.


> How do you envision the correctness of these solutions being judged?

By LLMs. I think it's possible for agents to infer whether the user was satisfied or not, at least with my usage pattern. For example if I end the discussion it's a good sign. If I ask follow up question that look like workarounds, it's a bad sign :-)

You could also simply prompt the users whether they were satisfied with the answer they received, possibly incentivizing them with StackOverflow-style gamification.


I'm sure the AI companies will continue to pirate textbooks and papers, like always.

Yeah, this is something I've been thinking about too. LLMs have basically profited from "stealing" (arguably) user-generated content from a time when there were no LLMs. In the LLM era there won't be a new Stack Overflow to train LLMs on going forward.

We're getting closer to Dead Internet Theory too where a lot of accounts, particularly on Twitter, are just LLMs. I imagine it's a huge problem on Reddit too. Just people farming karma or otherwise involved in influence campaigns or simply grifting to ad revenue.

So we're going to get to a point where the corpus we train LLMs on will itself just be filled with LLM slops. Self-reinforcing slop. Is that the future?


It's been studied,LLM that feed on its own data regress and it becomes very bad after a few generations.

It's happening here too, I saw dang hint that they're not even responding to a lot of questions about it anymore because of the volume of the problem.

If you browse with showdead on you'll be seeing a lot more of what look like reasonable comments greyed out.


Careful, you can't point out that the AI emperor has no clothes or you'll get called a Luddite.

Yes. Very.

How do you convince people to not want an instant answer? Even if SO didn’t result in so many “What have you tried?” responses and immediate closures, most people would still prefer instant feedback.

Pointing them to docs? Which is anyway what stack overflow answers did?

I wrote multiple answers to questions that weren't just "point to docs". And even when it is pointing to docs you are providing the reasoning as to why it works one way or another.

What docs? Who writes docs now that AIs answer everything?

Ever since the AI stuff started rolling around on coding i've seen MORE documentation, theres a big incentive to properly document your API endpoints so LLMs can figure it out from specs, and even when not documented the llms can also just read the code and figure it out directly (for libraries and similar). And at least in my experience they tend to document or write it down for future sessions too!

I know you're being facetious but there may well be docs. It's just that the same AI most likely wrote _them_, too.

Did anyone (person or competing LLM) bother to verify that they're correct, though? Who knows! Let the next generation of models worry about that.


Yeah, sorry, I guess I should be clearer that I'm rather sarcastic. My sad experience unfortunately shows that people less docs (or the docs are now hallucinated AI slop) instead of writing more of them.

I've heard this is now most of some CS jobs now. Just writing documentation for AI.

on the contrary, theres more of an incentive for apis to have docs for agent discovery. the docs / interfaces themselves can be auto-gened (stainless / mintlify)

> by being the one to discover the way of the future

This is my understanding too. The underlying assumption is that action leads to information, iterations lead to enlightenment. So from an org's point of view, tokenmaxxing means encouraging everyone to explore as much as they can. Of course, token volume should not be the only metric - tokenmaxxing is just a catchy phrase.


> action leads to information, iterations lead to enlightenment

So doing something (action) creates something new (more information), and iterating on that new information leads to the realization there is nothing new left to be learned with that information (enlightenment). Is how I'm interpreting that.


On the other hand, some companies are pushing the idea that engineers should build robust self-evaluating agent pipeline with human feedback in the loop so that agents write most of the production code. Creao's CEO said that they rearchitected their entire production systems in two weeks this January. He also claimed that their agents implemented so many features so fast that they had to wait their business development to catch up.

I wonder how we can evaluate these two options: using AI to 100X the output versus using AI to advance one's craft.

In the meantime, the productivity gain of AI is real. Case in point, An engineering org of Snowflake has met all its OKRs ahead of time in the first quarter for the time in the company's history. It had never happened, and usually meeting 70% of the planned OKR would be considered an achievement. I can imagine the stress of the engineers when they see such outcome.


I'm always hesitant of these claims. Sure, it's possible that AI really did help them achieve the same level of quality at 100x the pace. It's also possible it generated a huge tech debt that only passes the tests but hasn't planned for future maintainability, readability, and extensibility, and a year from now their entire process will grind to a halt.

I have a few people on my team who move 5-10x faster than others in writing code. They also generate 5-10x as many bugs and require that much more rework in the things that were shipped. They move fast and break things. Their code is almost malicious compliance in that it passes the tests or spec as given, while leaving glaring holes in things that weren't fully specified. A more careful developer would have asked questions, considered alternatives, and looked for ways to leverage existing solutions or plan for future work, but that takes time now and its benefits don't show up until later.

So while I don't immediately disbelieve that 10x+ speedups are possible with heavily AI-augmented flows, I am skeptical of any short term success stories until we have time to see the long term effects. We already know that cutting corners can save time in the short term only to cost us several times more in the long term.


Hopefully we can blend those two options together so it’s not a choice.

Personally I find being able lean on our heavily documented standards in /review gives me back time to dive into what I want to craft next.

Same with scheduling repetitive tasks an agent can do for me well once instructed well. I am freed up to do something else I want to focus actively on because I like it and want it to be great.

Now stress about OKRs and OKRS in general… that’s a different issue


Do you listen to anything while walking, or just listen to nothing while letting your mind clear itself?

Not OP, but it has to be a walk with no headphones for me. As I walk, thoughts seem to bubble up from my subconscious and present themselves for consideration. This doesn’t happen as often if I’m listening to music.

I decided to go offline for this summer. I got a dumb phone and a card for public transportation, instead of the app I'm using now.

Downtime from the algorithmic manipulation has been the breeding ground for my creativity and this is one more step to this direction.


I wish more people knew you can turn iPhones and Androids into dumbphones through MDM and other methods. It would save people money , you wouldn't have to sacrifice security, and they wouldn't complain about losing Google maps or Signal.

Result is no ability to install apps and no web browsing. It's really a smart, smartphone because you get the benefits of it being smart without becoming dumb through the distractions.


Anything I can remove, I can restore. So yes and no.

Few people have the willpower to stand against the addictive design, but I'm not one of them :D


You can use a password to make it so you can't restore. That's the difference with my methods.

There are various ways to store the password to allow some level of management. Give half of it to a friend, write it down, make it super long.


Why fight the system when you can just leave the system?

It would save people money, you wouldn't have to sacrifice security, and they wouldn't complain about losing Google maps or Signal.

Paper maps still work. What do you sacrifice in security in a dumb phone? A dumb phone is much cheaper. You can still call your friends.

Increasingly services want 2fa and other bullshit that only really plays nice with a modern smartphone. They don’t sell a lot of dumb phones fwiw. The network that your old one in the drawer ran on is shut down. The new “dumphones” are usually android phones designed for old people with poor eyesight and dexterity.

For 2FA, why would I want to use my phone? Certainly not SMS. YubiKey primarily, TOTP if necessary. Neither of which I need a phone for.

Most TOTP solutions are phone based, but you're right you can use them on any platform.

Some 2FA is app based, so that you'd need a phone for.

You wouldn't want to, but it's what 99% of people are herded into doing. TOTP is a lot more supported than hardware keys.


It's a mental thing too, the years of habit have built up such that for me smartphones are associated with distraction.

It's like deciding to quit smoking but using an empty cigarette pack to carry your credit cards. Sure, I'm not smoking, but every time I pay for something I have to squash the urge.


I deleted my browser and installed an app on my phone to block all apps except the ones that I have in an allowlist

So you have an article you can point to?

Cider9986 answered for Android, so I'll throw out a suggestion for iPhone.

Assistive Access on iPhone might be an option for people looking for something drastic. Turning it on is simple, but it's pretty brutal and a bit crude in some ways even compared to a feature phone. Your mileage will vary! It's something I often suggest, and never quite recommend.

https://support.apple.com/en-sg/guide/assistive-access-iphon...

You pick the apps you want access to, and the permissions each should have, set a password, and then when you turn Assistive Access on, the phone reboots into a very limited mode. You can have every app you want, but when I've played with it, I've still found it felt too limited for daily use. Maybe I wouldn't find that if I was at the point of buying a feature phone. I can't remember what frustrated me, except that I remember being pleasantly surprised by how much worked, and frustrated by some basic things.

As an example, I was impressed that I could turn on and off a VPN through an app, even though I couldn't see the status of it outside the app. On the other hand, the location permissions felt buggy, and the locations permission changes in Assisted Access mode seemed to mess with the settings in the normal mode too.


I didn't use an article, I just followed the principles and had an LLM do the android debug bridge commands.

Here is an article I found later which did the same thing as me.

(https://jordanherzstein.neocities.org/posts/adb_vanadium/)

For Android basically:

Live in user profile, keep owner profile with appstores. Push apps that are distractions free into user profile.

Use ADB to remove the built in browser because you can't just delete it or not install it because it's a system app. On GOS it's the only system app that is distracting, but I can imagine other phones might have others. Same principle, just remove it with ADB from the user profile.

Never install an app store in the user profile.

Owner profile password mitigation. You have a few options. Make it way too long to easily type and memorize it, write it down on paper and put it away in basement/attic/friends house, give it to a friend, give part of it to a friend(so they can't unlock the owner profile, only you can, but only if you ask them so huge friction).

Personally, I just have a super long passphrase memorized and that's enough too make the friction large enough. And it's really peaceful on the user profile.

Result. Without the owner password, I am in the user profile and I can't browse the web(HN) or install a distracting app like TikTok or install a new browser. If I want to update an app or manage the device or when the device restarts

Back when I was on iOS I used Apple Configurator which is Apple's MDM solution. You need a Mac it borrow one.

You remove Safari and disable installing apps. This is the guide I followed. Pretty sure your have to factory reset your phone first.

https://redd.it/1731ozp

So, to install new apps you have to connect the iPhone to the Mac and optionally add a password.

MDM is supported by Apple, uninstalling the browser is not recommended by GOS developers, but I haven't had any issues. Soon, GOS will support MDM, so hopefully that will be an even better solution.


I don’t walk but I run 60-120 min 4-5x a week and could not imagine doing so with headphones. Firmly believe we need time away from the constant stimulation of modern life.

I wish I could do the same, but the running(even at low pace like 6mph) is too taxing without something fun to listen to

I always find treadmill running to be as much of a mental workout staying focused as a physical one

Too taxing in what sense? Too boring? Too hard? If it’s the later, slow down to a brisk walk to build some stamina.

If it’s the former, start watching your surroundings. There’s a ton of things that are fun to watch.


Sounds like they’re using a treadmill, and yes this is about the most boring way possible to exercise

Mostly boring, but in upper zone 2 and sometimes zone 3 does not help. Yeah, I find it helpful to run outdoor. It’s particularly enjoyable to run in a trip because the routes will be unfamiliar

For several years I walked to and from the office, about 1.5 miles each way. Typically in the morning I would listen to a podcast or audiobook, and on the way home I would often continue thinking about whatever I had been trying to figure out at work. I found it useful.

> YouTube is eating itself from the inside out too

One thing that I really really hate Youtube for is that they don't allow users to turn off their shorts. You can choose to "reduce" Shorts for a given session, but they come back right next time.

That said, Youtube is tremendously valuable for its high-quality content. It's kinda like a restaurant. The service can be horrible. They decor can be hideous. But! I'm going back as long as the food is delicious.


You can go to Google Account > Data & Privacy. Then pause Youtube History. There will be no more feed on Youtube home screen. You will only see your /subscriptions feed. Little trick for a more peaceful life.

And that subscriptions screen has a row of all shorts across the top, and shorts trun up in searches and side bar related recommendations.

This while paying them for premium for 20 years.


There's a new setting rolling out in the YouTube app.

Go to settings > time management > shorts feed limit. Turn that setting on, and you can select how many minutes you limit to. There's now an option for "0 minutes".


This does nothing. It does not remove shorts.

On IOS/macos there's an app called "Unwatched for YouTube" which allows you to subscribe to channels via RSS (no need to login) and then you can turn shorts on/off per channel.

It's free for now but the developer has plans for some kind of subscription for premium features.

https://apps.apple.com/in/app/unwatched-for-youtube/id647728...



> but they come back right next time.

I never once did playables and each time asking them to dismiss them. I wrote down every time I was re-prompted for over a year:

March 19, 2025 - 8:31 PM

April 9 - 4:09 PM

April 24 - 8 AM

May 9 - 5:33 PM

May 20 - 2:07 PM

June 8 - 5:10 PM

July 9 - 6:59 PM

August 9 - 5:14 PM

September 8 - 8:45 PM

November 9 - 8:47 PM

December 9 - 8:48 PM

Jan 8, 2026 - 9:28 PM

Feb 7 — 11:11 PM

March 10 - 9:18 PM

April 10 - 1:10 AM

May 10 - 7:53 AM


Yeah the 'not being able to turn off shorts' is such a brazen, anti-user form of enshitification. Alongside not being able to hide threads in Instragram (can only hide for 30 days), and so many other examples. Like there is enough demand for this that there are literally browser extensions to block shorts.

I can see why youtube don't want you to disable; because shorts are "addictive" in a certain moorish way and letting you disable would lesson your expected youtube use time.

But it's such a wierd choice on a certain level right. Like "lets make our product objectively worse for users because (in the short term?) we'll make more money". It's the sort of choice that does't really exist in the "real" "normal" economy. Like you bake some bread, you wanna make it as good as possible, I buy it from you because you make good bread.

So anyway I get why they do it. I'm just a little surprised that in their calculations the gains to engagement from forcing shorts are worth the loss of user goodwill. And even like employee morale right. Like how would you feel about your job if you're having to do this stuff, deliberately and explicitly curtailing the choices of your users.

But yes I agree the content is great.


It’s a good decision. If an IDE can do everything that a CLI does and it surely can, then I fail to see the point of a CLI. It’s not like an IDE can’t emulate everything a CLI does but better, faster, and more interactive. It’s not like one does not need to read code either. Besides, what about session management? What about configuring agents, especially for multi-agent orchestration? The list can go on. The point is, IDE or GUI in general gives us optionality. Then, what’s wrong with that?

One may argue that Google’s Antigravity is clunky or cluttered or something worse, but that’s confusing organizational capability with principles.


Well, there is no IDE in antigravity 2.0

Ouch! I assumed too early

Which makes this part quite funny

"One may argue that Google’s Antigravity is clunky or cluttered or something worse, but that’s confusing organizational capability with principles"

The amount of free and biased good will Google gets here in HN is weird.


I don't want to use a full AI IDE, I've been using Gemini CLI a lot and it works better for my workflows to bring my own IDE and use Gemini CLI alongside it. My assumption is others are having annoyances for the same reason.

Curious: what motivates the Canadian government to implement such law? It's not like Canada wants to be a police state in anyway. On the contrary, Canadian government looks pretty chill most of the time, except maybe during the Covid era when they were hellbent on implementing the Covid policies. Or it's the same "for your own good and the state knows how to take care of you" kind of European shit?


> I don't think that AIs have become more trustworthy, the errors are just more subtle.

Honest question: what about the counter-argument that humans make subtle mistakes all the time, so why do we treat AI any differently?

A difference to me is that when we manually write code, we reason about the code carefully with a purpose. Yes we do make mistakes, but the mistakes are grounded in a certain range. In contrast, AI generated code creates errors that do not follow common sense. That said, I don't feel this differentiation is strong enough, and I don't have data to back it up.


One answer, as another person pointed out, is that LLM mistakes are just different. They are less explicable, less predictable, and therefore harder to spot. I can easily anticipate how an inexperienced engineer is going to mess up their first pull request for my project. I have no idea what an LLM might do. Worse, I know it might ace the first fifty pull requests and then make an absolutely mind-boggling mistake in the 51st one.

But another answer is that human autonomy is coupled to responsibility. For most line employees, if they mess up badly enough, it's first and foremost their problem. They're getting a bad performance review, getting fired, end up in court or even in prison. Because you bear responsibility for your actions, your boss doesn't have to watch what you're up to 24x7. Their career is typically not on the line unless they're deeply complicit in your misbehavior.

LLMs have no meaningful responsibility, so whoever is operating them is ultimately on the hook for what they do. It's a different dynamic. It's probably why most software engineers are not gonna get replaced by robots - your director or VP doesn't want to be liable for an agent that goes haywire - but it's also why the "oh, I have an army of 50 YOLO agents do the work while I'm browsing Reddit" is probably not a wise strategy for line employees.


> I can easily anticipate how an inexperienced engineer is going to mess up their first pull request for my project.

Isn’t this just because you have seen a lot of PRs from inexperienced engineers? People learn LLM behavior over time, too.


I'm pretty sure that I've seen more LLM mistakes than coworker mistakes at this point and I'm nowhere closer to enlightenment.


Humans can't make mistakes at the sheer scale that AI can.

Yes, as an engineer I make mistakes, but I could never make as many mistakes per day as an LLM can


Obviously, the measure isn’t mistakes per day, it’s mistakes per LOC. And that’s not the whole story either - AI self-corrects in addition to being corrected by the operator. If the operator’s committed bugs/LOC rate is as low as the unaugmented programmer’s bugs/LOC, you always choose the AI operator. If it’s higher, it might still be viable to choose them if you care about velocity more than correctness. I’m a slow, methodical programmer myself, but it’s not clear to me that I have a moat.


This is like having a coworker who's as skilled as you if not more skilled, but also an alien.

Their mental model doesn't map cleanly enough to yours, and so where for a human you'd have some way to follow their thought patterns and identify mistakes, here the alien makes mistakes that don't add up.

Like the alien has encyclopedic knowledge of op codes in some esoteric soviet MCU but sometimes forgets how to look for a function definition, says "It looks like the read tool failed, that's ok, I can just make a mock implementation and comment out the test for now."


Some of my favorite peer engineers work exactly like that

People used to like them and they used to be legends (even if not everyone liked them)

Notch, Woz, Linus and Geohot come to mind

The Metasploit creator Dean McNamee worked for me and he was just like that and a total monster at engineering hard tech products


No they don't because they have brains.

I have no strong idea why people can't accept that intelligence formed separately of a human brain can truly be alien: not in the hyperbolic sense of "that person is so unique it's like they're a different species", but "that thing does not have a brain, so it can have intelligence that is not human-like".

A human without a brain would die. An LLM doesn't have a brain and can do wonderous things.

It just does them in ways that require first accepting that there is no homo sapien thinks like an LLM.

We trained it on human language so often times it borrows our thought traces so to speak, but effective agentic systems form when you first erase your preconceived notions of how intelligence works and actually study this non-human intelligence and find new ways to apply it.

It's like the early days of agents when everyone thought if you just made an agent for each job role in a company and stuck them in a virtual office handing off work to each other it'd solve everything, but then Claude Code took off and showed that a simple brain dead loop could outperform that.

Now subagents almost always are task specific, not role specific.

I feel like we could leap ahead a decade if people could divorce "we use language, and it uses language so it is like us", but I think there's just something really challenging about that because it's never been true.

Nothing had this level of mastery over human language before that wasn't a human. And funnily enough, the first times we even came close (like Eliza) the same exact thing happened: so this seems like a persistent gap in how humans deal with non-humans using language.


I think these are reasonable questions but it assumes that everything is actually a black box instead of being treated as such.

Despite what the headlines say, these systems aren’t inscrutable.

We know how these things work and can build around and within and change parameters and activation functions etc…and actually use experience and science and guidance.

However those are not technical problems those are organizational social and quite frankly resource allocation problems.


I said the opposite of what your comment is replying to.

> but effective agentic systems form when you first erase your preconceived notions of how intelligence works and actually study this non-human intelligence and find new ways to apply it.

There's no reason you can't make good use of them and learn how to do it more reliably and predictably, it's just chasing those gains through a human intelligence-like model because they use human language leads to more false starts and local maxima than trying to understand stand them as their owb systems.

I don't think it should even be a particularly contentious point: we humans think differently based on the languages we learn and grew up with, what would you expect when you remove the entire common denominator of a human brain?


"I feel like we could leap ahead a decade if people could divorce "we use language, and it uses language so it is like us","

Or maybe just maybe... the thing should be much better designed around the human.

That's how personal computers made their way into homes. People like yourself are comical and can't understand how widespread adoption takes place to obtain value from what the thing intrinsically possesses.

Firms literally exist to take care of the hassle so that the person can get the value from the thing closer to the present - like hello...?


You quote me then start speaking about things completely unrelated to anything I said.

We can't choose if the LLM is like us unless you want to go back 10-20 years in time and choose a new direction for AI/ML.

We stumbled upon an architecture with mostly superficial similarities to how we think and learn, and instead focused on being able to throw more compute and more data at our models.

You're talking about ergonomics that exist at a completely different layer: even if you want to make LLM based products for humans, around humans, you have to accept it's not a human and it won't make mistakes like a human (even if the mistakes look human) -

If anything you're going to make something that burns most people if you just blindly pretend it's human-like: a great example being products that give users a false impression of LLM memory to hide the nitty gritty details.

In the early days ChatGPT would silently truncate the context window at some point and bullshit its way through recalling earlier parts of the conversation.

With compaction it does better, but still degrades noticeably.

If they'd exposed the concept of a context window to the user through top level primitives (like being able to manage what's important for example), maybe it'd have been a bit less clean of a product interface... but way more laypeople today would have a much better understanding of an LLM's very un-human equivalent to memory.

Instead we still give users lossy incomplete pictures of this all with the backends silently deciding when to compact and what information to discard. Most people using the tools don't know this because they're not being given an active role in the process.


Dealing with the alien coworkers has always been the job, that is what software is to most people.

Software developers get paid big money because they can speak alien, the only thing that is changing is the dialect.


Nope, I tried my best to be really detailed and already knew these replies would come flooding.

I'm an engineers engineer: I get the job isn't LOC but being able to communicate and translate meatspace into composable and robust sustems.

So when I mean an alien when I say an alien.

Not human.

Not in the cute "oh that guy just hears what everyone else hears and somehow interprets it entirely differently like he's from a different planet" alien way, but in the, "it is a different definition of intelligence derived from lacking wetware" alien way.

Intelligence is such multidimensional concept that all of humanity as varied as we are, can fit in a part of the space that has no overlap with an LLM.

-

Now none of that is saying it can't be incredibly useful, but 99% of the misuse and misunderstanding of LLMs stems from humans refusing to internalize that a form of intelligence can exist that uses their language but doesn't occupy the same "space" of thinking that we all operate in, no matter how weird or unqiue we think we are.


> Honest question: what about the counter-argument that humans make subtle mistakes all the time, so why do we treat AI any differently?

We're investing in the human getting better rather than paying $100 to Anthropic and hoping that's enough that they don't make the product worse.


Maybe it's just me, but isn't it exhausting that we have to do all kinds of work like a Shaman to combat the probabilistic nature of LLM, while knowing from the bottom of our heart that LLM can still screw up the code in the most creative way?


AI makes us believe that instead of working towards a goal, one can "win" that goal with a lucky prompt. AI replaces thinking with gambling, in other words, and it's very tempting to many.


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

Search: