[seertaak]

To my mind at least, it is different. I lean heavily on AI for both admin and coding tasks. I just filled out a multipage form to determine my alimony payments in Germany. Gemini was an absolute godsend, helping answer questions in, translate to English, draft explanations, emails requesting time extensions to the Jugendamt case worker.

This is super scary stuff for an ADHDer like me.

I have an idea for a programming language based on asymmetric multimethods and whitespace sensitive, Pratt-parsing powered syntax extensibility. Gemini and Claude are going to be instrumental in getting that done in a reasonable amount of time.

My daily todos are now being handled by NanoClaw.

These are already real products, it's not mere hype. Simply no comparison to blockchain or NFTs or the other tech mentioned. Is some of the press on AI overly optimistic? Sure.

But especially for someone who suffers from ADHD (and a lot of debilitating trauma and depression), and can't rely on their (transphobic) family for support -- it's literally the only source of help, however imperfect, which doesn't degrade me for having this affliction. It makes things much less scary and overwhelming, and I honestly don't know where I'd be without it.

[petterroea]

My empirical experience is that people with ADHD are more vulnerable to get addicted to LLMs due to the feeling of instant gratification. But when PRs take ages and 3 different people are reviewing, you are just making prompting a group effort. If you think meetings are a time waste multiplier you should watch LLM PRs.

For that reason, and my own experience with AI users being unaware of how bad of a job the LLM is doing (I've had to confront multiple people about their code quality suddenly dropping), if someone says they can rely on LLM I've learned to not trust them.

When I was younger if I had an idea for a project I would spend time thinking of a cool project name, creating a git repo, and designing an UI for my surely badass project. All easy stuff that gave me the feeling of progress. Then I would immediately lose interest when I realized the actual project idea was harder than that, and quit. This is the vibe I get from LLM use.

I pray you do not become the next HN user to be screwed over by over-trusting LLM when you have it fill out legal documents for you.

[seertaak]

I "pray" that you'll learn how work in an inclusive and non-toxic work environment.

What did I say? I lean on LLMs because I can't get help without being subjected to unnecessary degradation.

And what do you and others do? You immediately prove my point by saying things that amount to unnecessary degradation.

[petterroea]

I have many friends and loved ones with ADHD. It's very common in the IT industry, and probably >50% of people in the hacker spaces I frequent are neurodivergent in some way.

What I wrote is my empirical experience, but also what friends and loved ones tell me. I have friends with ADHD who have gone through the exact "wow I'm getting a lot done" -> "wow this is actually wasting a lot of time in hindsight" thing I described. If you think others lived experience is degrading to you it may be hitting a sore spot. What if I had ADHD? My friends with ADHD have the same opinion. Would you then say you were degraded by another person with ADHD that were offering their lived experience?

Maybe we live in very different countries but help has been good for everyone I know who got it. More want it the problem is money. You basically have to be suicidal to get public help, and private costs a fortune. It is a psychologists whole job to use their knowledge to help you self reflect and then act on it. It is uncomfortable, and I can understand why you may experience it as degrading. I don't know about the kind of help you've tried, though.

I hope you get the help you want.

[camillomiller]

This comment is scary. You don’t control these technologies, you are growing dependent on stilts that could disappear any moment.

[nicbou]

What if they’re just good for a while and then you go back to the old way?

[seertaak]

I'd be remiss to point out we went from "LLMs are vaporware" to "people are becoming slaves to their LLMs" awful quick.

> [I'm scared] you are growing dependent on stilts that could disappear any moment.

First, I do control the RTX3070 I own, and that can actually do a pretty decent job nowadays with some of the 3B parameter models.

Second, maybe if people like you showed as much concern for the fact that LGBT people can expect family violence as you do for Dr. Strangelove scenarios, then people like me wouldn't have to lean on LLMs so heavily.

Third, it's hilarious that your response to a comment pointing out how difficult it was to get help from another human without being degraded, was to degrade me by calling me an LLM junkie. Maybe you should be worried that Gemini appears to have more capacity for empathy and self-awareness than you.

Fourth, given that you show absolutely zero concern or willingness to help when it comes to the difficulties faced by LGBT people or ADHDers, my advice to you is to take your fears and shove them up your ass.

[camillomiller]

I am sorry, but literally fuck off. You don’t fucking know me. You don’t fucking know how much I do for the LGBTQ community, to which I belong, and honestly you just don’t fucking know shit about shit. Maybe you should start your journey by realizing how your problem is first and foremost the disgusting entitlement and victim mentality you show on this post. And also ask yourself why you seem to derive and perceive more empathy from stochastic sycophantic parrots than other human beings. But once again, let me reiterate: FUCK OFF.

[_zagj]

Peak HN: two people arguing over who's gayer in a thread about AI hype.

[camillomiller]

I’m arguing that I don’t need any deranged ai-addicted internet poster with a victim mentality to tell me how the fuck I live my life. Has nothing to do with my sexuality.

[hdgvhicv]

Holy pivoting Batman!

[Imustaskforhelp]

> First, I do control the RTX3070 I own, and that can actually do a pretty decent job nowadays with some of the 3B parameter models.

Quick question but what model are you exactly running with 3B parameter. The only decent model I can find which can compete sort of with Cloud models without costing a bank in GPU/RAM are the recently launched Qwen models (35A3B or 27B) which were released a week ago

> First, I do control the RTX3070 I own, and that can actually do a pretty decent job nowadays with some of the 3B parameter models.

My larger question to you is that even if it might not disappear in any moment, the fact of the matter still remains as if that its still a dependency. Is this dependency worth it? This is an open question and something I am still thinking.

> Third, it's hilarious that your response to a comment pointing out how difficult it was to get help from another human without being degraded, was to degrade me by calling me an LLM junkie. Maybe you should be worried that Gemini appears to have more capacity for empathy and self-awareness than you.

Gemini isn't real tho. It's still linear algebra with no regards to what it says or not. It's just trained on all the corpus data that Google can find and fine tuned to mimic it. By attaching real human qualities to Gemini, we dilute the value of those human qualities in those first place.

I don't necessarily know how "Humans" have treated you. They have treated me both good and bad but I am always more greatful to those who taught me or discussed with me things and helped me know something new. I very much feel like the same fine-tuning that I discussed earlier about models make those very agreeable and the chances of growth are rather limited.

> Fourth, given that you show absolutely zero concern or willingness to help when it comes to the difficulties faced by LGBT people or ADHDers, my advice to you is to take your fears and shove them up your ass.

Actually, You are a human as well so try to think it like this, I am sure you must've met both good and bad people and observed a few common characteristics of them. You are a human too and each second gives you a choice which can help you get either good or bad characteristics being better/worse each day.

Now my philosophy is to be good if not for yourself, then for others in the sense that you become the person that you wished could help you in your life and you can use that to actually help other people. This might be a little naive and practical nature sometimes might not follow this philosophy but yea.

So I want for you to reflect on what you wrote and think as if perhaps that might be a little too aggressive? and if that's what you want or not.

My or (our?) worry is that it feels like too big of a dependence on LLM which are fundamentally black boxes (yes, they are!), Humans can be bad but humans can be good too, I suggest even though it can be hard to have a good friend group (even if online) and talk with them about normal life issues.

Regarding, Coding, I would consider that there are some great people here on forums or Github or just about anywhere who are kind as well and can be helpful. Stackoverflow as an example had issues because of moderation problems which led to the community being hostile but to say that the whole of Software Engineering is such way might be wrong.

Speaking from personal experience, I may or may not have ADHD, I haven't diagnosed it yet but I definitely went into the AI=Producitivty rabbit hole especially more because I am a teen and I was in 9th/10th grade when ChatGPT came iirc. I knew basic python and knew the concepts of multiple languages and chatgpt felt hella addicting to be making websites in svelte all of a sudden where I can make one color button turn to another.

I wouldn't be lying if I say that I may not have learnt Coding effectively the way it was designed from its origin until quite recently. I was Vibe coding from the start and I have made quite some projects at the very least.

My observation is that its great for prototyping purposes but even after finally creating prototypes of most if not all the project ideas I ever had. I lost the motivation to continue and felt burn out. I did everything that I ever wanted to and made every project I thought yet the projects still felt hollow.

So, nowadays I am trying to focus more on studies for my college which can also act as a sort of recovery, to me it was also the fact that I was making these projects when I should've been studying in hindsight haha but I always just wanted to "prove" something (Yes I struggle with studies quite often but I wish to improve and I hope I can improve since I know from past that I can study often but its rather that I need my pure undirected focus on it which became hard for some time)

Recently, I went into a marriage of my own cousin. I found that to be much more fulfilling experience than expected. There is something about human experience both good or bad which can't be quantified.

I don't know what the future holds for me or you. But I wish you luck and hope this message helps ya. I personally realize that aside from prototyping which may be less meaningful than I previously thought at times, AI to me feels quite weak.

I think that for any product to really win, you might need true conviction in the product itself and at that point, the point of prototyping with AI or writing the code with AI to me becomes moot/redundant whereas AI is causing ram prices/storage to increase which is putting genuine projects out of luck as well. [This is one of the worst times to open a Cloud/VPS provider shop]

Perhaps I can understand AI use to get Open source tool when there were none or something but that to me seems like a cultural issue where Open source isn't funded so people are more likely to have it closed source to survive their likelihood but even that to me feels very moot point as there are some great open source projects as well who would appreciate each and every dollar that you donate to them, perhaps more so than a 200$ subscription of claude code as well which you might have to create the alternative to those in the first place as well.

My point still feels to me that it still feels hollow, I think you can find one of my other comments some days ago where I talk about this feeling of hollowness about AI projects as well which I can't help but feel relevant so many times. I am curious as to what you might think.

Have a nice day.

[unchar1]

The good thing is that local models are catching up very fast.

[artemonster]

Can you elaborate your choice about asymmetric multimethods? I also tinker with my PL and wanted to hear your reasonings and ideas

[seertaak]

Sure! First, here are references, in case you want to deep dive:

1. http://lucacardelli.name/Papers/Binary.pdf

2. https://www.researchgate.net/publication/221321423_Parasitic...

Second, asymmetric multimethods give something up: symmetry is a desirable property -- it's more faithful to mathematics, for instance. There's a priori no reason to preference the first argument over the second.

So why do I think they are promising?

1. You're not giving up that much. These are still real multimethods. The papers above show how these can still easily express things like multiplication of a band diagonal matrix with a sparse matrix. The first paper (which focuses purely on binary operators) points out it can handle set membership for arbitrary elements and sets.

2. Fidelity to mathematics is a fine thing, but it behooves us to remember we are designing a programming language. Programmers are already familiar with the notion that the receiver is special -- we even have a nice notation, UFCS, which makes this idea clear. (My language will certainly have UFCS.) So you're not asking the programmer to make a big conceptual leap to understand the mechanics of asymmetric multimethods.

3. The type checking of asymmetric multimethods is vastly simpler than symmetric multimethods. Your algorithm is essentially a sort among the various candidate multimethod instances. For symmetric multimethods, choosing which candidate multimethod "wins" requires PhD level techniques, and the algorithms can explode exponentially with the arity of the function. Not so with asymmetric multimethods: a "winner" can be determined argument by argument, from left to right. It's literally a lexicographical sort, with each step being totally trivial -- which multimethod has a more specific argument at that position (having eliminated all the candidates given the prior argument position). So type checking now has two desirable properties. First, it design principle espoused by Bjarne Stroustroup (my personal language designer "hero"): the compiler implementation should use well-known, straightforward techniques. (This is listed as a reason for choosing a nominal type system in Design And Evolution of C++ -- an excellent and depressing book to read. [Because anything you thought of, Bjarne already thought of in the 80s and 90s.]) Second, this algorithm has no polynomial or exponential explosion: it's fast as hell.

4. Aside from being faster and easier to implement, the asymmetry also "settles" ambiguities which would exist if you adopted symmetric multimethods. This is a real problem in languages, like Julia, with symmetric multimethods. The implementers of that language resort to heuristics, both to avoid undesired ambiguities, and explosions in compile times. I anticipate that library implementers will be able to leverage this facility for disambiguation, in a manner similar to (but not quite the same) as C++ distinguishes between forward and random access iterators using empty marker types as the last argument. So while technically being a disadvantage, I think it will actually be a useful device -- precisely because the type checking mechanism is so predictable.

5. This predictability also makes the job of the programmer easier: they can form an intuition of which candidate method will be selected much more readily in the case of asymmetric multimethods than symmetric ones. You already know the trick the compiler is using: it's just double-dispatch, the trick used for "hit tests" of shapes against each other. Only here, it can be extended to more than two arguments, and of course, the compiler writes the overloads for you. (And it won't actually write overloads, it will do what I said above: form a lexicographical sort over the set of multimethods, and lower this into a set of tables which can be traversed dynamically, or when the types are concrete, the compiler can leverage monomorphize -- the series of "if arg1 extends Tk" etc. is done in the compiler instead of at runtime. (But it's the same data structure.)

6. It's basically impossible to do separate compilation using symmetric multimethods. With asymmetric multimethods, it's trivial. To form an intuition, simply remember that double-dispatch can easily be done using separate compilation. Separate compilation is mentioned as a feature in both the cited papers. This is, in my view, a huge advantage. I admit, this I haven't quite figured out generics will fit into this -- at least if you follow C++'s approach, you'll have to give up some aspects of separate compilation. My bet is that this won't matter so much; the type checking ought to be so much faster that even when a template needs to be instantiated at a callsite, the faster and simpler algorithm will mean the user experience will still be very good -- certainly faster than C++ (which uses a symmetric algorithm for type checking of function overloads).

To go a bit more into my "vision" -- the papers were written during a time when object-orientation was the dominant paradigm. I'd like to relax this somewhat: instead of classes, there will only be structs. And there won't be instance methods, everything will be a multimethods. So instead of the multimethods being "encapsulated" in their classes, they'll be encapsulated in the module in which they're defined. I'll adopt the Python approach where everything is public, so you need to worry about accessibility. Together with UFCS, this means there is no "privileging" of the writer of a library. It's not like in C++ or Java, where only the writer of the library can leverage the succinct dot notation to access frequently used methods. An extension can import a library, write a multimethod providing new functionality, and that can be used -- using the exact same notation as the methods of the library itself. (I always sigh when I read languages, having made the mistake of distinguishing between free functions and instance methods, "fix" the problem that you can only extend a library from the outside using free functions -- which have a less convenient syntax -- by adding yet another type of function, an "extension function. In my language, there are only structs and functions -- it has the same simplicity as Zig and C in this sense, only my functions are multimethods.)

Together with my ideas for how the parser will work, I think this language will offer -- much like Julia -- attractive opportunities to extend libraries -- and compose libraries that weren't designed to work "together".

And yeah, Claude Code and Gemini are going to implement it. Probably in Python first, just for initial testing, and then they'll port it to C++ (or possibly self-host).

[artemonster]

Thanks for elaborated reply, both papers Ive seen too. I have mostly same views, but I really dislike that there is no clean solution for binary methods, i.e. add( float, int), where symmetric add(int, float) ends up being a boilerplate. Also I think in asymmetric case its hard to handle dispatch when it has failed to produce method when looking in first argument. i.e. dispatching "collide" with Asteroid, Ship, if collider method is found in Ship, how to bind "this", where does Asteroid is bound. Anyways, good luck with your experiments!

[tomlue]

"This time is different" has been correct for every major technological shift in history. Electricity was different. Antibiotics were different. Semiconductors were different.

Gen AI reached 39% adoption in two years (internet took 5, PCs took 12). Enterprise spend went from $1.7B to $37B since 2023. Hyperscalers are spending $650B this year on AI infra and are supply-constrained, not demand-constrained. There is no technology in history with these curves.

The real debate isn't whether AI is transformative. It's whether current investment levels are proportionate to the transformation. That's a much harder and more interesting question than reflexively citing a phrase that pattern-matches to past bubbles.

[bigstrat2003]

> The real debate isn't whether AI is transformative.

No, the debate is very much whether AI is transformative. You don't get to smuggle your viewpoint as an assumption as if there was consensus on this point. There isn't consensus at all.

[selridge]

No one is smuggling this in. The debate is over. It's transformative. We're in the midst of transformation.

[bandrami]

It's really not over. Somebody has to actually put something into production with it first.

[jwittmayer]

Implying that nobody has put AI generated code into production yet?

[bandrami]

Stuff that's going into production now (actual production, not startup MVP production) would have been being written just before Claude Code came out, so pretty much by definition no. There's some copilot-style assisted stuff in the wild, I guess? But not really more of it than pre-copilot so the productivity argument kind of falls through there.

[adaml_623]

I feel like you might only be convinced when an AI powered robot rolls up to you and asks, "Bandrami, are you convinced that AI is transformative yet?"

[bandrami]

Robots have been able to do that for decades now

[bitwize]

No, it is over. Compare today to even two years ago.

[usrnm]

I put AI assisted code in production every day, what are you talking about? At this point I don't even doubt I'm going to lose the job eventually, the question is only whether or not I will be able to pay my mortgage off first.

[cogman10]

The problem is in the middle of such a change it's hard to recognize if this is a real change or if this is another Wankel motor.

Plenty a visual programming language has tried to toot their own horns as being the next transformative change in everything, and they are mostly just obscure DSLs at this point.

The other issue is nobody knows what the future will actually look like and they'll often be wrong with their predictions. For example, with the rise of robotics, plenty of 1950s scifi thought it was just logical that androids and smart mechanic arms would be developed next year. I mean, you can find cartoons where people envisioned smart hands giving people a clean shave. (Sounds like the making of a scifi horror novel :D Sweeney Todd scifi redux)

I think AI is here to stay. At very least it seems to have practical value in software development. That won't be erased anytime soon. Claims beyond that, though, need a lot more evidence to support them. Right now it feels like people just shoving AI into 1000 places hoping that they can find an new industry like software dev.

[jibal]

I once owned a Maxda RX2 ... my second car, IIRC. The Wankel motor wasn't revolutionary, but it was pretty good.

[radlad]

> Plenty a visual programming language has tried to toot their own horns as being the next transformative change in everything, and they are mostly just obscure DSLs at this point.

But how many of your non-nerdy friends were talking about them, let alone using them daily?

[bojan]

The practical value is there, if they managed to keep the price at the current levels or lower.

But if they don't and if I have to think twice about how much every request's going to cost, the cost-benefit analysis will look differently fast.

[cogman10]

Yeah that's another rub. The current price is basically there in the hopes that in the future they can find revenue streams to maintain their current pace.

But even if the big companies ultimately go belly up, I think the open models are good enough that we'll likely see pretty cheap AI available for a while, even if it's not as good as the STOA when the bankruptcies roll through.

[rsynnott]

> Sounds like the making of a scifi horror novel :D

See ‘Service Model’. YMMV on whether you consider it horror.

[fragmede]

The four technologies I look at are 3D televisions, VR, tablets, and the electric car. 3D televisions and VR have yet to find their moment. Judging tablets by the Apple Newton and electric cars by the EV1, this time is different turns out to be the correct model looking at the iPad and Tesla, but not for 3d televisions or VR (yet). So, it could be, but my time machine is as good as yours (mine goes 1 minute per minute, and only forwards, reverse is broken right now.), so unless you've got money on it, we'll just have to wait and see where it goes.

[alpaca128]

> Gen AI reached 39% adoption in two years (internet took 5, PCs took 12)

You're comparing a service that mostly costs a free account registration and is harder to avoid than to use, with devices that cost thousands of dollars in the early days.

[tomlue]

That is a fair point. You could look at enterprise adoption though, also very high, and not cheap at all.

[thesz]

  > 39% adoption in two years (internet took 5, PCs took 12).

Adjust for connectivity and see whether it is different (from pure hype) this time.

[legulere]

There's another perspective you can see in the comparison with the dot com boom. The web is here to stay, but a lot of ideas from the beginning didn't work out and a lot of companies turned bankrupt.

[inigyou]

The original concept of the web, hyperlinked documents originating from high-quality institutions, is pretty much dead. Now we have an application platform that happens to have adopted some similar protocols and is 99% slop

[hdgvhicv]

It wasn’t surprise me if a lot of AI companies go bankrupt.

However some will survive, and there will be far more bankruptcy and downsizing in the industries replaced

[spidersouris]

> Gen AI reached 39% adoption in two years

Source?

[tomlue]

https://www.stlouisfed.org/on-the-economy/2024/sep/rapid-ado...

[Jensson]

So about 10%, using it less than once per day means you didn't find it useful for most tasks.

[hdgvhicv]

Just like the PC. Or the internet.

In 1995 how many people used the internet in their daily work, of those that did how many was it a curiosity that maybe supplemented their existing business practice (sending a memo via email rather than post for example). Large companies were using large computer mainframes but the majority of employers - the SMEs - weren’t.

By 2005 it massively shifted, and AI seems to be coming faster than the internet and computers in general.

By 2015 non intenet companies were going the way of the dodo. How many travel agents were there per 100k in 1995 compared to 2015?

[shimman]

My boss never had to threaten me to use a computer, unlike the current LLM mandates across corporate America.

[shimman]

Also add in that these adoption rates are being enforced via threats of firing by bosses of workers. It's hardly something organic, there's a reason why the LLM companies are chasing lucrative corporate welfare contracts because consumers have soundly rejected this nonsense.

[201984]

Yeah, what's counting as "adoption" here?