Well, the conditional in this if statement doesn't hold.

Yes, pharmaceuticals are open access in much of the developing world, but it has not happened responsibly. For example, Carbapenem-resistant bacteria are 20 times as common in India as they are in the U.S [1]

I really don't like this characterization of medical resource stewardship as "infantilization" because it implies some sort of elitism amongst doctors, when it's exactly the opposite. It's a system of checks and balances that limits the power afforded to any one person, no matter how smart they think they are. In a US hospital setting, doctors do not have 100% control over antibiotics. An antibiotic stewardship pharmacist or infectious disease specialist will deny and/or cancel antibiotics left and right, even if the prescribing doctor is chief of their department or the CMO.

[1] https://www.fic.nih.gov/News/GlobalHealthMatters/may-june-20...

[1] https://github.com/google/dawn/blob/40cf7fd7bc06f871fc5e4823...

[2] https://github.com/protocolbuffers/protobuf/blob/c964e143d97...

[3] https://github.com/facebook/react-native/blob/77b3a8bdd6164b...

[4] https://github.com/NVIDIA/cccl/blob/07fef970a33ae120c8ff2a9e...

However the adoption rates of newer C++ features are in fact new are way lower. From what I see lots of projects still use the language as C with Classes, basically, and that ain't going to change any time soon. The GP nailed it - C++ is adding a lot of esoteric stuff that very few people actually need or want.

There is a reason why C++17 is the best we can currently hope for in what concerns portable code, given the actual support across industry compilers, and company project guidelines.

Many embedded shops might still be discussing between adopting C++11 or C++14.

A friend of mine who works in primary care has a policy of not prescribing semaglutide unless a patient has tried at least three months of diet/exercise first. She now has a long list of patients who decide to screw that and self-refer to a specialist who'd write them the script. I don't work in primary care myself, but this more or less matches my own experience - 95% of people do not want to put in the effort of changing diet, exercise, and other lifestyle habits. They want a quick fix.

In the words of Ronnie Coleman, "Everybody wants to be a bodybuilder, but nobody wants to lift no heavy-ass weights."

So I really do think the popularity of semaglutide is a bottom up phenomenon. There is has been huge consumer demand for a weight loss drug for decades, and pharma is only now meeting that demand.

> doctors having intimate financial relationships with pharma is already a public secret

A small number of elite doctors have intimate relationships with pharma. The other 99% who are prescribing semaglutide do not profit from it. In fact, it's the other way around. You can't bill for the act of prescribing medications, but there are dedicated billing codes for counseling a patient on diet/lifestyle modification.

I lost a bunch of weight when I moved to Dublin and started walking everywhere. Then I moved to the Netherlands and lost even more, and got stronger. I didn't particularly try, I just lived somewhere where I used my own body to get around and the "normal" portion sizes were actually somewhat sane (a bag of Doritos here is maybe a third the size of one in the US).

I was 280 pounds when I lived in suburban California. I'm 180 pounds now.

Build bike lanes and public transport, put schools where kids can walk to them, and maybe don't sell a thousand calories of chips in a personal-sized and marketed bag, and it should get better.

As an added bonus you get a lot less air pollution (and noise pollution, etc.) and less people being run over!

GLP-1 pharma intervention is breadboarding the human, next step is a permanent fix. Adversarial bioengineering, if you will.

Regarding your proposed gene therapy, would this gene therapy be available to anyone? Could it be refused? Is it reversible? Who would pay for it? The beauty of The Culture is that it is not monolithic, has no border, has godlike AI that takes care of all the problems, and can be abandoned or joined by anyone.

I know what you meant, but according to a comment here once, European city streets actually have worse air pollution than US ones because Europe went hard for diesel engines (to reduce CO2 emissions) for its cars, and diesel produces a lot of very small carbon particles, which turned out to be more harmful than most people suspected back when the diesel decision was made.

[1]: https://www.amsterdam.nl/en/traffic-transport/low-emission-z...

Of course, Volkswagen lying about their diesels and killing people in the aggregate didn’t help.

What we need is car-free or low-car cities, not just different cars.

Popular EV's aren't significantly heavier than ICE vehicles. A Tesla model 3 and a BMW model 3 are both about 3500 pounds. Range is a huge selling feature for EV's and weight has a significant impact on range, so there is strong evolutionary pressure to lighten EV's. Crappy EV's like the Hummer are heavy; but popular EV's like Tesla's and Hyundai's have weight comparable to normal cars so that they can have impressive ranges without too many expensive and heavy batteries.

EV's increased tire wire is due to their massive torque at low range. If you want your tires to last, just don't drive like a hooligan. We got 70,000 km out of our first set of EV tires.

Tire particulates are an "all cars a bad" issue. EV's are not significantly worse than petrol vehicles vis a vis tire particulates. Their lack of exhaust particulates means that an EV produces about ~half the total number of particulates.

Even so, I still find that the fewer cars there are, the easier it is for me (and my kids) to walk and bike places safely, helping us be healthier (and not reliant on drugs like Ozempic) which was the real point of my comment.

Everyone is directly responsible for their own health and nobody else is. The insanity of this statement is baffling.

If the genetic / hormonal explanation is right, it would mean the "effort" is not the same for everyone. So expecting people on the wrong side of the equation to not go for a solution to right the scales demonstrate a lack of empathy.

Would you call cochlear implants a quick fix for people who have problem hearing?

There's nothing with wanting a quick fix to make it easier and less unpleasant. I'd happily use a provably safe one. The only issue is that currently that fix carries unknown risks.

this is a crisis.

at this point this is like going to a 3rd world country and telling them if you dont want all the problems that come with not having money to just get rich.

like its possible, my brother did it, most my family did it, we all have the capability to do it. it maybe harder for some due to physical limitations or mental ones. so why dont you? dont hate being poor and not getting all the women and luxuries?

74% is sick and its infectious. it curates mental problems that make it even harder to overcome. 74% means there are too many factors that are contributing to this epidemic. if modern technology is able to help get society on the right track i dont know why anyone should be.. for lack of a better word fat shamed haha.

of course we need to be careful and thoughtful. im not even sure though if it will be available for most people anytime soon. i hope so though, iv never been fat per say. maybe a bit more lbs then id like but i do understand how much it helps every part of your life being at a satisfying weight. this could be the greatest cure for depression through medicine ever imo.

i live in one of the fattest cities in america. when i go outside i swear to you sometimes i can go a day without seeing a single person of normal weight. besides the few homeless in my area.

Rephrasing the famous quote, it's easy to lose weight — I've done it four times already, ranging from 95 kg to 60 kg over the course of the last 15 years. It's much more difficult to maintain healthy weight over the long term — the longest period I managed that was only three years. Even if you have all the information and know which problems obesity leads to, it's difficult to keep yourself from getting back into unhealthy territory unless you're willing to spend your entire life counting calories and tracking weight religiously.

It's something with the brain, it's insatiable and won't leave you alone until you stuff yourself to the point that you physically can't eat any longer. The last time I gained weight (from 68 kg to around 88 kg) I did it eating only healthy food in unhealthy amounts.

If I can get magic fix (like semaglutide) that would allow me not to think about food any longer and maintain healthy weight over the rest of my life, that would be great.

It depends on your habits. My habits are generally good enough that I don't need to count calories. Which is mostly luck, not an achievement of mine, btw.

> The last time I gained weight (from 68 kg to around 88 kg) I did it eating only healthy food in unhealthy amounts.

First, there isn't even clear consensus what's healthy and what isn't. But even if there was, I doubt you only ate healthy food. It is very hard to gain much fat eating a lot of vegetables and some fish and eggs.

Of course it's possible your diet is great. My experience with fat people has been that their diet was a whole lot worse than they claimed...

In absolute, changing diet is not inherently hard : it's pretty easy to eat a balanced diet and get pleasure from it.

But it's not enough to just say "eat better", "do more sports" : people can't change the way the lived their whole life if you don't educate and help them to make the change. The education on this topic is so poor that we are still confusing balanced diet and low-calorie diet.

And then there is the elephant in the room which is the refined sugar : for some dubious reason, most scientists agree that refined sugar acts like a hard drug but nobody officially wants to call it a drug. Recent research are saying that the addictive effects of refined sugar are akin to heroin. That's a serious issue then. What it means is that a doctor asking his patient to eat less sugar / eat a balanced diet without further guidance or help is just like saying to an heroin addict that he should stop.

I'm relatively overweight and pretty self educated on the topic but I still decided to work with a nutritionist to change my diet and we are talking about months if not years of follow-up. Eating habits are largely automated by our biology : willpower and education are nothing against an hypoglycemia induced sugar craving.

Like drugs, most people need help and guidance on the long term to change those habits. But what's even worse is that unlike drugs, unhealthy products are basically everywhere in your supermarket, your TV ads, your billboards, your friends lifestyle, and even worse, profoundly ingrained in the culture. You have to fight the "drug" but also the whole capitalist world around you.

But even a balanced diet can let you put weight if you eat too much.

We know genetics disorder like Prader-Willis can cause insatiable appetite. Even with a "balanced diet" those affected by it will end-up obese.

Now the current research on GLP-1 hint at many people having a harder time feeling satiated, not to the Prader-Willis point but still worse than what is considered normal. For people who don't have the problem it is easy to think they just have more discipline / willpower and that's why they're thin. I'd like them to imagine what would happen if after eating 2 or 3 pizza slices they'd still feel like their stomach is empty. What if it's not a one time occurrence but all day every day from the moment they develop a conscience to their death bed. A little like drug addiction but you need some drug to live so it is legal and you can buy it everywhere. Still think you'd have the discipline to not go for the whole pizza?

Being obese with a balanced diet is absolutely not the same thing than being obese because you eat too much "bad" calories. Obese or not, a balanced diet gives you everything your body needs to function properly, be able to move and for your mind to be clear.

The thing is, obesity is a really, really recent trend in humanity history, so recent that a lot of countries are still not concerned by the phenomena (but it's changing). It's pretty certain that _something_ in the environment changed recently and made the humans obese.

People who blames other people willpower just don't understand that if obesity is in an uptrend, it means that people who were previously healthy are becoming obese for some reason. Something new pressurizes humans to become obese and the only question is not if you are concerned but when will the pressure be high enough.

I think it's refined sugar. Plus a bonus of sedentary lifestyles.

Even that is the generous interpretation. Imagine instead of just their regular (or lack of) hunger signals, people had an alarm constantly ringing in the head telling them to eat all the time, that was only ever silenced if they over ate. The persistence of it pushes it in the direction of mental torture.

Yes, technically it can be overcome, but it is a huge tax on your life.

The truth is we can never really know what someone else's inner experience is.

This ozempic study seems to suggest something similar.

Self-discipline is a noble concept, and probably important in the grand scheme of things, but sometimes medicine and technology outpaces what is naturally possible, which inevitably results in a moral backlash.

Well yes obviously that is difficult to do. Our society (American society, that is) is about “me” and convenience.

Nobody seems to want to say what is obvious to me: ban junk food. Sugary white bread, Cheez Its, Oreos, candy, ice cream, highly processed frozen burritos, frozen pizza, Little Debbie’s, potato chips , all of it.

It is crazy to me the C levels of poison food companies do not rot in jail for the rest of their lives for knowingly selling poison and lying about it. It is crazy to me how a group of people can conspire to harm the citizens of a nation in the name of profit and not receive any consequences. It’s crazy to me the governments of that nation actually enjoy it and don’t care.

The data however shows that diet and exercise fails for most people though. I don't see it as entirely irrational to go strait to the more effective drugs.

Many are unlucky to have unforgiving disorders.

This is easily seen in T2D which is mostly diet related and fixable.

I decided to fire up GPT-4o again today to see if maybe things have gotten better over the past few months.

I asked GPT to write code to render a triangle using Vulkan (a 3D graphics API). There are about 1000 tutorials on this that are almost certainly in GPT-4's training data. I gave GPT two small twists so it's not a simple case of copy/paste: I asked it 1) to apply a texture to the triangle and 2) to keep all the code in a single function. (Most tutorials break the code up into about a dozen functions, but each of these functions are called only once, so it should be trivial to inline them.)

Within the first ten lines, the code is already completely nonfunctional:

GPT-4o declares a pointer (VkPhysicalDevice) that is uninitialized. It queries the number of graphics devices on the host machine. A human being would allocate a buffer with that number of elements and store the reference in the pointer. GPT-4o just ignores the result. Completely ignores it. So the function call was just for fun, I guess? It then tries to copy an entire array of VkPhysicalDevice_T objects into this uninitialized pointer. So that's a guaranteed memory access violation right off the bat.

Some basic things are fine, but once you get into specialised things, everything gets just terribly wrong and weird. I can't even put it into words, I see people saying how they are 10x more productive (I'd like to see actual numbers and proof for this), but I just don't see how. Maybe we're working on very custom stuff, or very specific things, but all of these tools seem to give very deep or confident answers that are just plain wrong and shallow. Just yesterday I used GPT 4o for some basic help with Puppet, and the examples it printed, even though I would say it's quite basic, were just wrong, but in the sense of having to debug it for 2 hours just to figure out how ridiculous the error was.

I fear the fact that people will end up releasing unsafe, insecure and simply wrong code every day, code that they never debug and not even understand, that maybe works for a basic set of details, but once the real world hits it, it will fail like those self driving cars driving full speed into a trailer that has the same color as the road or sky.

However, my main job is trying to come up with "elegant" architectures for complex business logic that interacts with an existing large code base. AI is just completely out of its depth in such cases due to lack of context but also lack of source material to draw from. Even unit tests only work with the most basic of cases, most of the time the setup is so complex it just produces garbage.

I've also had very little luck getting it to write performant code. I almost have to feed it the techniques or algorithms before it attempts to write such code, and even then it's usually wrong or not as efficient as it could be.

That code is probably buggy, slow, poorly architected, very verbose, and has logical issues where the examples and your needs dont exact match.

Generally, the longer the snippet you want your LLM to generate, the more likely its going to go off the rails.

I think for some positions this can get you 90% of the code done. For me this usually means I can get started very fast on a new problem, but the last remaining "10%" actually takes significantly longer and more effort to integrate because I dont understand the other 90% off the top :)

Right, that's surely the main job of nearly every experienced developer. It's really cool that LLMs can generate code for isolated tasks, but they can barely even begin to do the hard work, and that seems very unlikely to change in the foreseeable future.

I just tried the latest Cursor + Sonnet and it failed in every task. The problem is that there is no way to understand the code without either complete understanding of the domain and the intents or running it in some context.

Telecom and media domains in particular are well documented in specs and studied in forum discussions. I am sure they are part of the training data, because I can get most answers if asked directly. So far the LLMs fail to reason about anything useful for me.

The Jetbrains thing is rather rudely incompetent, it consistently insists on suggestions that use variables and fragments that are supposed to be replaced by what I'm writing and also much more complex than what I actually need. I suffered through at least a hundred mistaken tabbed out shitty suggestions before I disconnected it.

However, I use copilot autocomplete consistently. That makes me much more productive. It’s just really good autocomplete.

As an experienced developer I’d put my productivity improvement at 2-3x. It’s huge but not 10x. I’m limited by my decision speed, I need to decide what I want the code to do, AI can’t help with that - it can only do the “how”.

Far from introducing more bugs, using Copilot frees some mental cycles for me to be more aware of the code I’m writing.

So it will be with inexperienced people coding with LLMs.

On stuff that's a bit obscure, but not really (like your Vulkan example), ChatGPT tends to write 60-95% correct code, that's flawed in the exact ways a noob wouldn't be able to fix.

In this case, nicking code from Github seems to fix the issue, even if I need to adapt it a bit.

Then comes the licensing issue. Often, when searching for an obscure topic, the code ChatGPT generates is very close to what's found on Github, but said code often comes with a non-permissive license, unlike what the AI generates.

I think this is a feature for a lot of people. ChatGPT can launder code so they don't have to care about licenses.

"They clearly aren't using the right model!"

"It's obvious they don't know how to prompt, or they would see the value."

"Maybe it can't do that today, but GPT-5 is just around the corner."

I feel more and more that people have just decided that this is a technology that will do everything you can imagine, and no evidence to the contrary will change their priors.

Oh and I expect it to be free, I ain't paying for this just like I wasn't paying for stackoverflow.

Finally I hope than in few years I will be able to just "sudo apt-get install llm llm-javascript llm-cooking llm-trivia llm-jokes" and it will all run locally on my low end computer and when I report bug, six months later it will be fixed when I update OS.

The same applies to AI. The old learning material is gone, your interaction is now the new learning material and ground truth.

PS: Hourly rates for sw engineers: Range:€11 - €213 - so one hour on stackoverflow, searching and sub-querying resolving problems costs you or your employer up to 213€. It really depends on what you have negotiated.

How about DDG AI Chat? https://duckduckgo.com/?q=DuckDuckGo+AI+Chat&ia=chat&duckai=...

It's like having an unusually fast-but-clumsy intern, except interns learn the ropes fast and understand context.

I have some stand tests for LLMs: write a web app version of tetris, write a fluid dynamics simulation, etc., and these regularly fail (I must try them again on 4o).

But also, I have examples of them succeeding wildly, writing a web based painting app just from prompting — sure, even with that success it's bad code, but it's still done the thing.

As there are plenty of examples to confirm what we already believe, it's very easy to get stuck, with nay-sayers and enthusiasts equally unaware of the opposite examples.

I welcome people picking apart chats that I link to. it's not that I believe that LLMs are magic and refuse to adjust my model of how good these things are and aren't, but when people don't give specific evidence is hard to actually move the conversation forwards.

because yeah, these things are plenty stupid and have to be tricked into doing things sometimes (which is stupid, but here we are). they're also pretty amazing but like any hammer, not everything is a nail.

You're like a gambling addict who thinks he's smarter than everyone else

By the way - I think AI or ML whatever has some valid uses right now. but mostly in image processing domain - so like recognizing shapes in some bounded domain OK yea. Generative image - NOT bad but theres always this "AI GLOW" to each image. Something is always off. Its some neat tools but a race to the bottom and mostly users want to generate explicit content lets be real. and they will become increasingly more creative and obtuse to get around the guards. nothing is stopping you from entering the * industry and making tons off money. but that industry is always doing good.

a friend recently suggested to use AI to generate generic icons for my game. Thats a really good use case. but does that radically change the current economy?

[BTW GENERIC STUFF ONLY UNTILL I could hire someone because i prefer that experience way more. you can get more interesting results. 4 eyes are better than 2.

I am an experienced programmer, but I find myself (for the first time) doing a deep-dive in SQL and specifically building code that will run against multiple SQL engines.

I fed my list of engines into AI. As I'm discovering the really weird corners of SQL I asking the AI to compare one db against the other. My prompts are usually no more than 3 or 4 words.

It gives me quick helpful answers highlighting where things are the same and where they are different. I can then follow up in the specific docs if necessary (now that I know the function name.)

Personally I'm somewhat anti-hype, I'll let others rave about "changing the world". But I have found it a useful tool - not so much for "writing my code" but for acting as my tutor as I learn new things. (I'm using it for more than just SQL or computers now.)

I'm not sure it "changes thr economy" - but it can certainly change individual lives. Some jobs will go away. Others might be easier to do. It might make it easier to learn new skills.

The 2nd agency I called them through phone. No ai. But it is excellent cause they do async processing. so you reserve a slot and they call you back. i don't care if those are not answered urgently. Because I just want to talk to a human.

Customer service is hard because it has to filter out 90% noise, 9% fairly straightforward tasks, and the <1% of complex issues that need to be sent to the right human.

But equally it doesn't prove, or even assert, the opposite. A bicycle may be bad at cross-country road trips, but that doesn't make it a bad choice for some other situations.

Hence my earlier comment - I (and I suspect others) are finding it useful for some tasks. That is not to imply it is good at all tasks.

Is it over-hyped? Of course yes. Welcome to IT where every new thing is over-hyped all the time.

If you don't know the language at all it's dangerous because you may not understand the proposed program (and of course if you're an expert you don't need it at all).

But llms won't help to find solutions to a general, still unspecified problem.

For example, I can ask the LLM things like "What are the most common mistakes when using the Vulkan API to render a triangle with a texture?" and I'll very rapidly learn something about working with an API that I don't have deep understanding of, and I might not find a specific tutorial article about.

As another example, if I'm an experienced OpenGL programmer, I can ask directly "what's the Vulkan equivalent of this OpenGL API call?" and get quite good results back, most of the time.

So I'm asking questions where an 80% answer is still very valuable, and it's much faster than searching for documentation and doing a lot of comparison and headscratching, and it works well enough even when there's no specific article I could find in a variety of searches.

Anything better that the technology gets from here just makes things even easier yet!

For example, just now my NAS stopped working because the boot device went offline. So I got to thinking about writing a simple syslog server. I've never looked at the syslog protocol before, and I've never done any low-level TCP/UDP work in C# yet.

So I asked ChatGPT to generate some code[1], and while the result is not perfect it's certainly better than nothing, and would save me time to get going.

As another example, a friend who's not very technical wanted to make an Arduino circuit to perform some automated experiment. He's dabbled with programing and can modify code, but struggles to get going. Again just for kicks, I asked ChatGPT and it provided a very nice starting point[2].

For exploratory stuff like this, it seems to provide a nice alternative to searching and piecing together the bits. Revolutionary is a quite loaded word, but it's certainly not just a slight improvement on what we had before LLMs and instead feels like a quantum leap.

[1]: https://chatgpt.com/share/f4343939-74f1-404d-bfac-b903525f61... (modified, see reply)

[2]: https://chatgpt.com/share/fc764e73-f01f-4a7c-ab58-f43da3e077...

I've also assembled a Kubernetes cluster overnight, despite not knowing much about Kubernetes before, and I ran the kubectl files ChatGPT made for me past some devops folks, and it passed the smell test.

I consider much of coding to be a magic spellcasting tutorial - we do conceptually simple things, and the difficult lies in figuring out how to use novel libraries and get them to do what you want.

Edit: After checking out the Arduino sketch, I'd take issue with all the floating point calculations in there - most microcontrollers don't have FPUs, and the performance is awful on 8-bit AVRs. It's not great on Cortex M3s either as all this stuff is done in software, and each FP operation is like a hundred cycles.

I'd definitely try to rephrase the issue with integer math. It might work, but no self-respecting embedded dev would write like this.

Exactly, it's a very nice alternative to searching the web and discovering new stuff.

> most microcontrollers don't have FPUs, and the performance is awful on 8-bit AVRs.

I used to think like you. But then I realized the Atmega 328p is running at 16 MHz, so even hundreds of . As you can see here[1], it can do 94k double-precision FLOPS, more than enough for simple sketches like this. This jives with the benchmarks I did several years ago.

Sure if I was writing some tight control loop or similar I wouldn't use floating point.

[1]: https://kreier.github.io/benchmark/LinpackDP/

It's like I have some light fixtures in my attic that are connected with wires twisted together and covered with electrical tape - they certainly work and have done for a decade, but its still not right and I wouldn't recommend anyone do it this way.

Its using the old format where the Program.cs file has an actual class, whereas as of .NET 6 thats not required.

You said barebones, but for any real server you would want to use the generic host https://learn.microsoft.com/en-us/dotnet/core/extensions/gen... which gets you a lot of the boilerplate and enables you program to be wrapped in a windows or systemd service.

Finally, parsing can be simplified since ASCII is a proper subset of UTF-8, you can just parse the entire string as UTF-8. IMHO I am disappointed that the AI didn't point that out.

True, I intentionally said barebones as I wanted a minimal example. I asked it to modify the code to use the generic host, and updated the chat link (so refresh). Keep in mind this is the free ChatGPT, but I still think it did reasonably good. The example compiles as-is, and is very close to functional. I've not used the generic host stuff before either, so again this would save me time searching and piecing together code.

> Finally, parsing can be simplified since ASCII is a proper subset of UTF-8, you can just parse the entire string as UTF-8.

I don't think that would work, because the free-form text message part at the end must contain a BOM if it's UTF-8 encoded, according to the specification. AFAIK you can't have the BOM in the middle of a string.

I just don’t understand where the hell is this magical LLM capable of generating flawless files or even entire projects that many people are talking about. I rarely accept a large block of LLM-generated code without close inspection, and I’ve ended up with a subtle bug that I wouldn’t have written myself at least ~5 times now. Unless you don’t give a shit about code quality, error handling, proper logging, and subtle bugs, you shouldn’t run LLM-generated stuff without painstakingly reading and fixing everything. Or maybe there really is a magical LLM somewhere.

The way the OP here was talking about sonnet being above way above chatgtp in this case, it could be true. Google probably has the largest Go codebases on search to train the AI on higher quiality inputs. Go is a simpler language with less variation over time compared to something like .net also adding to its corner.

ive always been the type of person to use the right language for the each use case. For the last 10+ years ive primarly been building cross platform apps that target every common OS. So these "ai" tools like phind.com give me a universal API interface and generate code which is equiv to a SO answer. They are the ability of an outsourced junior dev who you would never let push code to prod, that doesnt have the language barrier but retains the fake dergee overheads ;)

It seems to work best if I start with something very simple, and then layer on instructions ("now make it do X").

Where I have found it saves me time is in having to look up syntax or "gotchas" which I would otherwise search StackOverflow for. But as far as "writing code" -- it still feels a long way from that.

For example:

* I need a simple bash script for file manipulation or some simple tasks like setting up a project (example: download a secret from AWS SSM, check if an executable exist, if it doesn't write instructions on how to install it on most popular systems etc)

* I need a simple HTTP API, nothing fancy, maybe some simple database usage, maybe running some commands, simple error handling

* I need a YAML file for Kubernetes. I say what I need and usually, it gets most of it right

* I want an Ansible task for some simple thing. Ansible is quite verbose, so it's often saving me time

* I have a Kubernetes YAML file, but I want to manage it in terraform - I'll then ask to convert YAML to a terraform entry (and in general converting between formats is nice, cause even if you have only a piece of what you want to convert, LLMs will most of the time get it right)

* Surprisingly, it often gets openssl and ffmpeg commands right - something I always have to google anyway, especially openssl certificates generation or manipulation

* Giving it a function I wrote and writing test samples after providing a list of what it should test (and asking if it can come up with more, but sadly it rarely does generate anything useful on top of what I suggest)

A friend, whose SQL knowledge is minimal, used an LLM to query data from a database over a couple of tables. Yes, after a lot of trial and error he (most probably) got the correct data, however the only one being able to read the query is the LLM itself. It's full of coalesce, subselects that repeat the same joins again and again.

LLM will do a lot for you, but I really hate this "this will [already did] solve everything". No, it did not and no, because it's quality is those of a junior dev, at max.

Echo chamber/bias I guess. I know many, many seniors with big pay checks working for big companies who are vastly worse than sonnet at the moment. Juniors stand literally no chance unless very talented.

I don't know what companies pay for and I don't care, because if we go by that, every definition of every word is arguable (because there's everywhere someone out of the range of the definition)

And I think it'd be extremely easy to convince oneself of this. Look at where 'AI' was 5 years ago, look at where it is today and then try to imagine where it will be in another 5 years. Of course you have to completely blind yourself to the fact that the acceleration has clearly sharply stalled out, but humans are really good at cognitive dissonance, especially when your perception of your future depends on it.

And there's also the point that even though I'm extremely critical of LLMs in general, they have absolutely 'transformed' my workflow in that natural language search of documentation is really useful. Being able to describe a desired API, but in an overly broad way that a search engine can't really pick up on, but that an LLM [often] can, is just quite handy. On the other hand, this is more a condemnation of search engine tech being frozen 20 years in the past than it is about an imminent LLM revolution.

I would rather say that the innovation happens in spikes. I see no slowing down whatsoever.

Especially if it's a question that's hard to Google, like "I remember there is more than one way to split an array in this language, list them". This saves me minutes every day.

But it's especially helpful if you are working on projects outside your own domain where you are a newbie.

That is very believable, the people who say it made them super productive is a bit less believable.

Cursor is a purpose-built IDE for software development. The Cursor team has put a lot of research and sweat into providing the used LLMs (also from OpenAI/Anthropic) with:

- the right parts of your code

- relevant code/dependency documentation

- and, importantly, the right prompts.

to successfully complete coding tasks. It's an apple and oranges situation.

I work in 9 different projects now and I would say that around 80% of functional code comes from Sonnet (like GP) for these projects. These are not (all) trivial either; there is a very niche (for banking) key/value store written in Go for instance which has a lot of edge cases etc, all the plumbing (x,err = etc aka stuff people find annoying) comes from sonnet and works one-shot. A lot of business logic comes from sonnet too; it works but usually needs a little tweaking to make it correct.

Tests are all done by Sonnet. I think 80% is low balling it on Go code really.

We have a lot of complex code generator stuff and DSLs in TS which also works well often. Sometimes it gets some edge cases wrong, but either re-prompting with more details or fixing it ourselves, will do it. At a fraction of the time/money of what a fully human team would deliver.

I wrote a 3d editor for fun with Sonnet in a day.

I have terrible results with gpt/copilot (copilot is good for editing instead of complete files/functions; chatgpt is not good much compared with sonnet); it doesn't get close at all; it simply keeps giving me the same code over and over again when I say it's wrong; it hardcodes things specifically asked to make flexible etc. Not sure why the difference is so massive all of sudden.

Note: I use the sonnet API, not the web interface, but same for gpt so...

However once you step outside JS or Python, the models are essentially useless. Comprehension of pointer semantics? You wish. Anything with Lisp outside its training corpus of homework assignments? LOL. Editing burden quickly exceeds any possible speed-up.

But, I agree with your sentiment that asking it to do stuff like that often doesn’t work. I’ve found that what it _can_ do is stuff like “here’s a Model object, write a query to fetch it with the schema I told you about ages ago”. It might not give perfect results, but I know how to write that query and it’s faster to edit Claude’s output than it is to write it from scratch.

Fwiw, I've had some helpful successful prompts here and there, and in some very narrow scopes I'll get something usable, like parsing JSON or scaffolding some test cases, which is real saved time, but I stopped thinking about these tools long ago.

To get real value out of something like your example, I'd be using it as a back and forth to help me understand how some concepts work or write example questions I can drill on my own, but nothing where precision matters

There are also some more gotchas, like the generated code using a slightly different package versions than the installed ones.

Same, it can't even fix an xcode memory leak bug in a simple app. It will keep trying and breaking it non-stop. Garbage

If you define "productive" as writing a simple CRUD web application that your 13-year-old cousin could write between two gaming sessions, then you'll consider LLMs as sacred monsters.

Snake oil vendors always had great appeal over people who didn't know better.

AI is great for me, but it is more like a junior developer you are pairing with than a replacement.

Like simple python scripts for Home Assistant it just nails first go.

Give Anthropic a shot (its even better via the API console.anthropic.com/workbench).

OpenAI is yesterdays news.

Why? I see it like querying a database of human knowledge. I wouldn't expect a SQL database to infer information it's never seen before, why would I expect an LLM to do so?

I use it where I know a solution exists but I'm stumped on the syntax or how to implement it in an unfamiliar environment, or I want to know what could have caused a bug based on others' experience etc.

here's a chat for a uc and LCD chip that I picked at random (and got the name wrong for) (and didn't want raspberry pi code for so it stopped it short on that response)

https://chatgpt.com/share/2004ac32-b08b-43d7-b762-91543d656a...

Would be really interesting if anyone had blog posts on their actual workflow with LLMs, in case there's something I'm doing different.

When you are familiar with LLMs, then a question from someone who doesn't use AI is very obvious. It's the same feeling you have when you roll your eyes and say "you could have googled that in 10 seconds".

It's either explaining code where you don't even know the lingo for or what the question could be. Or touching code with a framework you never used. Or tedious tasks like convert parts of text into code or json. Or sometimes your mind is stuck or drifts off. Ask AI for an idea to get the ball rolling again.

Yes, discovering what works and what doesn't is tedious and slower then "just doing it yourself". Like switching IDEs. But if you found a handful of usecases that solve your problems, it is very refreshing.

I saw an LLM demo at one point where it was asked to write FFT and add unit tests for it which really drove this point home for me.

A programmer is a nicer term for a code monkey. You ask them to write FFT and they'll code it. All problems can be solved with mode code. They can edit code, but on the whole it's more just to add more code. LLMs are actually pretty good at this job, in my experience. And this job is important, not all tasks can be engineered thoroughly. But this job has its scaling limits.

A software engineer is not about coding per se, it's about designing software. It's all about designing the right code, not more code. Work smarter, not harder, for scale. You ask them to write FFT and they'll find a way to depend on it from a common implementation so they don't have to maintain an independent implementation. I've personally found LLMs very bad at this type of work, the same way you and others relying to you describe it. (Ok, maybe FFT is overly simple, I'm sure an LLM can import that for you. But you get the idea.) LLMs have statistical confidence, not intellectual confidence. But software engineering generally works with code too complex for pure statistical confidence.

No offense to the LLM fans here, but I strongly suspect most of them are closer to the programmer category of work. An important job, but one more easily automated away by LLMs (or better software engineering long-term). And we can see this by how a lot of programming has been outsourced for decades to cheap labor in third-world countries: it's a simpler type of job. That plus the people biased because their jobs and egos depend on LLMs succeeding.

I’ve found that it’s sometimes amazing and sometimes wastes a lot of my time. A few times it’s really come up with a good insight I hadn’t considered because the conversation has woken up some non-obvious combination. I use ChatGPT, Claude, Perplexity and one or two IDE tools.

It seems especially strong with Python but a bit medium with Swift.

After that it goes completely off the rails by trying to issue draw commands before binding a graphics pipeline, which is both illogical and illegal. After a lot of prodding, I did manage to get it to bind a graphics pipeline, but it forgot about the texture.

So Claude Sonnet is definitely better than GPT-4o, but it still feels raw, like a game of whack-a-mole where I can get it to fix a mistake, but it reintroduces an old one. I also have to be the one offering the expertise. I can prompt it to fix the issues because I know exactly what the issues are. If I was using this to try to fill in for a gap in my knowledge, I would be stuck when I ran the code and it crashed - I would have no idea where to go next.

Update: Took about 50 min of experimenting, but I did get Claude to generate code that doesn't have any obvious defects on first inspection, although it cut off about halfway through because of the generation limit. That's the best result that I've seen from an LLM yet. But that's after about a dozen very broken broken programs, and again, I think the domain expertise here is key in order to be able to reprompt and correct.

One example that I am still using: I wanted to generate a random DICOM file with specific types of garbage in it to use as input for some unit tests, and Claude was able to generate some Python that grabs some random DICOM tags and shoves vaguely plausible garbage data into them, such that it is a valid but nonsensical DICOM dataset. This is not hard, but it's a lot faster to ask Claude to do it.

That would be ideal, but getting 195 countries on the same page on cybercrime just isn't going to happen. As it is we have multiple countries where the government actively sponsors internet scammers.

Also probably not quite as hard as you make out in practice, as 27 of those countries share membership of a super-national government whose specific purpose is to make trade easier by streamlining legislative differences between them, and that block plus two other single countries constitute about 60% of the world's GDP.

Though random small-time scammers may regardless have a hard time selling in China an app made outside, and vice-versa, as even just the language barrier can be quite intense — I've played a game where the decently resourced western publisher didn't notice (or noticed but didn't fix) the fact that the "choose your name" box only had room for one character, and thus I was the only player who had an interesting name like '狐' rather than 'M'.

You don't even need an extradition treaty. Just require some collateral for selling outside your own country. Then people selling within their own country (i.e. probably most people) don't have to post anything but are automatically in the same jurisdiction as the victim. Whereas scammers from other countries forfeit their collateral, the amount of which is set based on the amount of scams coming out of their country.

Corporations are only efficient when they have corporate responsibilities. Corporations do only efficiency and can select customers. Government must guarantee equality and rights for everyone. Even criminals are citizens with rights.

Let me demonstrate. This is how Google would do it:

(1) Algorithm captures 90% of all criminals (it's a good algorithm)

(2) 5% (1/20) change that flagged account is a criminal (95% false positives)

(3) 0.01% (1/10000) of all accounts are flagged.

There are around 246 million unique Google users in the US. Closing just 24,600 accounts removes 90% of criminals. 90% change of capture is a good deterrent.

Google also removes 23,370 innocent accounts.

GOOGLE DOES THIS ALREADY. It's efficient and well-run (actual numbers may vary) but also brutal and unjust. Legacy government institutions do their job better than Google would.

With Google you can only post complaint to HN and wish that someone working in Google notices, or that there is enough publicity to shame them.

I think Americans have forgotten what the government is. It's the job of the people to fix and constantly maintain the government because if they don't own it, someone else will.

It seems like Reaganism never left America. People from right to left have adopted the talking points and the attitude. If there are problems in the government, it's suddenly all bullshit and good for nothing. Talk in a passive voice where the government just happens.

Rust has UB-as-a-feature too. They could have eliminated UB from the language entirely, but they chose not to (for very valid reasons in my opinion).

UB is a set of contracts that you as the author agree to never violate. In return, you get faster code under the assumption that you never actually encounter a UB condition. If you violate those contracts in Rust and actually encounter UB, that's a a bug, that's a fault in the code. If you violate those contracts in C++, that's a bug, that's a fault in the code. This is the same in both languages.

It's true that Rust UB can only arise from unsafe blocks, but it is not limited to unsafe blocks. Rust UB has "spooky action at a distance" the same way C++ UB does. In other words, you can write UB free code in Rust, but if any third party code encounters UB (including the standard library), your safe code is now potentially infected by UB as well. This is also the same in both languages.

There are good reasons to favor Rust's flavor of UB over C++'s, but I keep seeing these same incorrect arguments getting repeated everywhere, which is frustrating.

This is correct, and it's hard to teach, and I agree that a lot of folks get it wrong. (Here's my attempt: https://jacko.io/safety_and_soundness.html.) But I think this comment is understating how big of a difference this makes:

1. Rust has a large, powerful safe subset, which includes lots of real-world programs. Unsafe code is an advanced topic, and beginners don't need to learn about it to start getting their work done. Beginners can contribute to big projects without touching the unsafe parts (as you clarified, that means the module privacy boundaries that include unsafe code, not just the unsafe blocks), and reviewers don't need to be paranoid about every line.

2. A lot of real-world unsafe Rust is easy to audit, because you can grep for `unsafe` in a big codebase and zoom right to the parts you need to look at. Again, as you pointed out, those blocks might not be the whole story, and you do need to read what they're doing to see how much code they "infect". But an experienced Rust programmer can audit a well-written codebase in minutes. It's not always that smooth of course, but it's a totally different world that that's even possible.

Tell me what I wrote that was incorrect. I called them UB bugs in the standard library. If they were trivial bugs that caused some defined-behavior logic bug while used outside of the standard library then it wouldn’t rise to the level of being called an UB bug.

That's the part that's incorrect. That, plus the implication that UB is a bug in Rust, but not in C++. As I said, the existence of UB is a feature in both languages and actually encountering UB is a bug in both languages. You can play with the semantics of the word "feature" but I don't think it's possible to find a definition that captures C++ UB and excludes Rust UB without falling into a double standard. Unfortunately double standards on UB are pretty common in conversations about C++ and Rust.

Do you think UB-as-feature is something that someone would honestly describe C or C++ as? It’s a pretty demeaning way of framing things. Indeed it’s a tongue-in-cheek remark, a vhimsical exaggeration/description of the by-default UB of those languages which was added to the end of the completely factual description of the role that finding UB in the Safe Rust subset of the standard library of Rust serves.

Of course one cannot, from the Rust Side so to speak, use tongue in cheek, off-hand remarks in these discussions; one must painstakingly add footnotes and caveats, list and mention every trivial fact like “you can get UB in unsafe blocks”[1] or else you have a “double standard”.

[1] Obligatory footnote: even though all participants in the discussion clearly knows this already.

Yes. That's how I describe it. That's also how Ralf Jung (long time Rust contributor and one of the main people behind Miri) describes UB in both Rust and C++ (although he says C++ overdoes it) [1]

The thing I edited out of my comment was "motte and bailey fallacy" because after reflecting a bit I thought it was unfair. But now you're actually trying to retroactively reframe as a joke.

[1] https://blog.sigplan.org/2021/11/18/undefined-behavior-deser...

Okay. Then I was wrong about that.

> The thing I edited out of my comment was "motte and bailey fallacy" because after reflecting a bit I thought it was unfair. But now you're actually trying to retroactively reframe as a joke.

What a coincidence. I had written on a post-it note that you were going to pull out an Internet Fallacy. (I guess it’s more about rhetoric.)

I guess you’ve never seen someone explain after the fact that they were being tongue in cheek (it’s not a joke, it’s an exaggeration)? Because jokes, sarcastic remarks are always clearly labelled and unambiguous? Okay then. I guess it was a Motte and Bailey.

I tried a toy example with GCC [1], Clang [2], and MSVC [3], and none of them emit bounds checks with basic optimization flags.

[1] https://godbolt.org/z/W5e3n5oWM

[2] https://godbolt.org/z/Pe8nPPvEd

[3] https://godbolt.org/z/YTdv3nabn

You can attach about a hundred asterisks to that.

If anything, I think self the failure to hit L5 driving after billions of dollars and millions of man hours invested is probably reflective of how automatic C to Rust translation will go. We'll cruise 90% of the way, but the last 10% will prove insurmountable with current technology.

Think about the number of C programs in the wild that rely on compiler-specific or libc-specific or platform-specific behavior, or even undefined behavior plus the dumb luck of a certain brittle combination of {compiler version} ∩ {libc version} ∩ {linker version} ∩ {build flags} emitting workable machine code. There's a huge chunk of C software where there's not enough context within the source itself (or even source plus build scripts) to understand the behavior. It's not even clear that this is a solvable problem in the abstract.

None of that is to say that DARPA shouldn't fund this. Research isn't always about finding an industrial strength end product; the knowledge and expertise gained along the way is important too.

In 10 (or 2 or 50 or X) years when Chevy, Ford, and others are rolling out cheap self driving this argument stops working. The important thing is that this argument stops working with no change in how hard C to Rust conversion is.

We really should be looking at the specifics of both problems. What makes computer language translation hard? Why is driving hard? One needs to be correct while inferring intent and possibly reformulating code to meet new restrictions. The other needs to be able to make snap judgments and in realtime avoid hitting things even if it just means stopping to prefer safety over motion. One problem can be solved piecewise without significant regard to time and the other solved in realtime as it happens without producing unsafe output.

These problems really aren't analogous.

I think you picked self driving cars just because it is a big and only partially solved problem. One could just as easily pick a big solved problem or a big unstarted problem and formulate equally bad arguments.

I am not saying this problem is easy, just that it seems solvable with sufficient effort.

I'd put money on the solutions to said problems looking largely the same though - big ass machine learning models.

My prediction is that a tool like copilot (but specialized to this domain) will do the bulk of source code conversions, with a really smart human coming behind to validate.

The director orders the use of the tool when the dev team got downsized (and the two most-seniors left for greener pastures just after that). Validation is in the "extensive" tests anyway, we have those, right, so the new intern shall have a look, make it all work (fudge the tests where possible and remove the persistently failing ones as they've probably been always broken). The salesman said it comes from the DOA or DOD or something. If the spooks can do it so can we.

Which are things that took 20 or 50 years longer than expected in some cases.

> I think you picked self driving cars just because it is a big and only partially solved problem. One could just as easily pick a big solved problem or a big unstarted problem and formulate equally bad arguments.

But C to Rust translation is a big and only partially solved problem.

Even if it's "can handle well 90%" of the transition for any project, this is still interesting. Unlike cars on the road, most code transition project out there doesn't need to be 100% fine to provide some useful value.

Just getting something vaguely Rust shaped which can compile is the first step in overcoming the inertia to leave the program in its current language.

There is real utility in slowly fleshing out the number of transforms in a tool like c2rust that can recognise high-level constructs in C code and produce idiomatic safe equivalents in rust

It'll be great for "pure" functions. For the grimey parts of the world, funcs taking pointer args and returning pointers, for things that access and modify global data without locks, for threaded code with implicit (and undocumented) locking, the tool would add most value. If it can. Even only by saying "this code looks grimey. here's why. A bit of FFI will also be thrown in because it links against 100 libraries. I suggest changes along those lines ... use one of the 2000000 hint flags to pick-your-evil".

> You can attach about a hundred asterisks to that.

Not in San Francisco. There are about 300 Waymo cars safely driving in one of the most difficult urban environments around (think steep hills, fog, construction, crazy traffic, crazy drivers, crazier pedestrians). Five years ago this was "someday" science-fiction. Frankly I trust them much more then human drivers and envision a future utopia where human drivers are banned from urban centers.

To get back on topic, I don't think automatic programming language translation is nearly as hard, especially since we have a deterministic model of the machines it runs on. I can see a possible approach where AI systems take the assembler code of a C++ program, then translate that into Rust, or anything else. Can they get 100% accuracy and bit-for-bit compatibility on output? I would not bet against it.

Getting identical outputs on safe executions, catching any unsafe behavior (at translation-time or run-time), and producing efficient, maintainable code all at once is a million times harder.