Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I don't care how well your "AI" works

fokus.cool

Post reply on HN
Print view View original

261–270 of 808 posts

Re: I don't care how well your "AI" works

#261

Post by scarface_74 »

cjbgkagh wrote: post #194
spicyusername wrote: post #115

Earlier quoted context omitted.

> assign work to an LLM This is just not happening anywhere around me. I don't know why it keeps getting repeated in every one of these discussions. Every software engineer I know is using LLM tools, but every team around me is still hiring new developers. Zero firing is happening in any circle near me due to LLMs. LLMs can not do unsupervised work, period. They do not replace developers. They replace Stack Overflow…

Around 80% of my work is easy while the remaining 20% is very hard. At this stage the hard stuff is far outside the capability of LLM but the easy stuff is very much within its capabilities. I used to hire contractors to help with that 80% work but now I use LLMs instead. It’s far cheaper, better quality, and zero hassle. That’s 3 junior / mid level jobs that are gone now. Since the hard stuff is combinatorial comple…

Exactly this, I lead cloud consulting + app dev projects. Before I would have staffed my projects with at least me leading it and doing the project management + stakeholder meetings and some of the work and bringing a couple of others in to do some of the grunt work. Now with Gen AI even just using ChatGPT and feeding it a lot of context - diagrams I put together, statements of work, etc - I can do it all myself without having to go through the coordination effort of working with two other people.

On the other hand, when I was staffed to lead a project that did have another senior developer who is one level below me, I tried to split up the actual work but it became such a coordination nightmare once we started refining the project because he could just use Claude code and it would make all of the modifications needed for a feature from the front end work, to the backend APIs, to the Terraform and the deployment scripts.

I would have actually slowed him down.

Re: I don't care how well your "AI" works

#262

Post by hvb2 »

markbnj wrote: post #201

I think the dangers that LLMs pose to the ability of engineers to earn a living is overstated, while at the same time the superpowers that they hand us don't seem to get much discussion. When I was starting out in the 80's I had to prowl dial-up BBSs or order expensive books and manuals to find out how to do something. I once paid IBM $140 for a manual on the VGA interface so I could answer a question. The turn aroun…

> So I guess my advice is: if you're starting out in this business worry less about LLMs replacing you and more about how to efficiently use that global expert on everything that is sitting on your shoulder.

There's a real danger in that they use so many resources though. Both in the physical world (electricity, raw materials, water etc.) as well as in a financial sense.

All the money spent on AI will not go to your other promising idea. There's a real opportunity cost there. I can't imagine that, at this point, good ideas go without funding because they're not AI.

Re: I don't care how well your "AI" works

#263

Post by jillesvangurp »

JimmaDaRustla wrote: post #192

We don't care that you don't care

Harsh but fair. In short, some people are upset about change happening to them. They think it's unfair and that they deserve better. Maybe that's true. But unfair things happen to lots of people all the time. And ultimately people move on, mostly. There's a futility to being very emotional about it.

I don't get all the whining of people about having to adapt. That's a constant in our industry and always has been. If what you were doing was so easy that it fell victim to the first generation of AI tools that are doing a decent enough job of it, then maybe what you were doing was a bit Ground Hog day to begin with. I've certainly been involved with a lot of projects where a lot of the work felt that way. Customer wants a web app thing with a log in flow and a this and a that. 99% of that stuff is kind of very predictable. That's why agentic coding tools are so good at this stuff. But lets be honest, it was kind of low value stuff to begin with. And it's nice that people over-payed for that for a while but it was never going to be forever.

There's still plenty of stuff these tools are less good at. It gets progressively harder if you are integrating lots of different niche things or doing some non standard/non trivial things. And even those things where it does a decent job, it still requires good judgment and expertise to 1) be able to even ask for the right thing and then 2) judge if what comes back is fit for purpose.

There's plenty of work out there supporting companies with decades of legacy software that are not going to be throwing away everything they have overnight. Leveling up their UIs with AI powered features, cross integrating a lot of stuff, etc. is going to generate lots of work and business. And most companies are very poorly equipped to do that in house even if they have access to agentic coding tools.

For me AI is actually generating more work, not less. I'm now taking on bigger things that were previously impossible to take on without involving more people. I have about 10x more things I want to do than I have bandwidth for. I have to take decisions about doing things the stupid old way because it's better/faster or attempting to generate some code. All new tools do is accelerate the pace and raise the ambition levels. That too is nothing new in our industry. Things that were hard are now easy, so we do more of them and find yet harder things to do next. We're not about to run out of hard things to do any time soon.

Adapting is hard. Not everyone will manage. Some people might burn out doing that or change career. And some people are in denial or angry about that. And you can't really expect others to loose a lot of sleep over this. Whether that's unfair or not doesn't really matter.

Re: I don't care how well your "AI" works

#264

Post by Aurornis »

zwnow wrote: post #249
Aurornis wrote: post #234

Earlier quoted context omitted.

> I am actually less productive when using LLMs because now I have to read another entities code and be able to judge wether this fits my current business problem or not. You don’t have to let the LLM write code for you. They’re very useful as a smart search engine for your code base, a smart refactoring tool, a suggestion generator, and many other ways. I rarely have LLMs write code for me from scratch that I have t…

> But you have to actually learn how to use them. This probably is the issue for me, I am simply not willing to do so. To me the whole AI thing is extremely dystopian so even on a professional level I feel repulsed by it. We had an AWS and a Cloudflare outage recently, which has shown that maybe it isn't a great idea to rely on a few companies for a single _thing_. Integrating LLMs and using all these tools is just a…

> This probably is the issue for me, I am simply not willing to do so.

Thanks for being honest at least. So many HN arguments start as a desire to hate something and then try to bridge that into something that feels like a takedown of the merits of that thing. I think a lot of the HN LLM hate comes from people who simply want to hate LLMs.

> We had an AWS and a Cloudflare outage recently, which has shown that maybe it isn't a great idea to rely on a few companies for a single _thing_. Integrating LLMs and using all these tools is just another bridge people depend on at some point.

For an experienced dev using LLMs as another tool, an LLM outage isn’t a problem. You just continue coding.

It’s on the level of Google going down so you have to use another search engine or try to remember the URL for something yourself.

The main LLM players are also easy to switch between. I jump between Anthropic, Google, and OpenAI almost month to month to try things out. I could have subscriptions to all 3 at the same time and it would still be cheap.

I think this point is overblown. It’s not a true team dependency like when GitHub stop working a few days back.

Re: I don't care how well your "AI" works

#265

Post by bgwalter »

easterncalculus wrote: post #244

> I find it particularly disillusioning to realize how deep the LLM brainworm is able to eat itself even into progressive hacker circles. That's the thing, hacker circles didn't always have this 'progressive' luddite mentality. This is the culture that replaced hacker culture. I don't like AI, generally. I am skeptical of corporate influence, I doubt AI 2027 and so-called 'AGI'. I'm certain we'll be "five years away"…

Being anti "AI" has nothing to do with being progressive. Historically, hackers have always rejected bloated tools, especially those that are not under their control and that spy on them and build dossiers like ChatGPT.

Hackers have historically derided any website generators or tools like ColdFusion[tm] or VisualStudio[tm] for that matter.

It is relatively new that some corporate owned "open" source developers use things like VSCode and have no issues with all their actions being tracked and surveilled by their corporate masters.

Please do no co-opt the term "hacker".

Re: I don't care how well your "AI" works

#266

Post by thwarted »

idle_zealot wrote: post #70
flir wrote: post #3

Earlier quoted context omitted.

Now apply that thinking to computers. Or levers. I've seen the argument that computers let us prop up and even scale governmental systems that would have long since collapsed under their own weight if they’d remained manual more than once. I'm not sure I buy it, but computation undoubtedly shapes society. The author does seem quite keen on computers, but they've been "getting rid of the free-willed human in the loop"…

> The author does seem quite keen on computers, but they've been "getting rid of the free-willed human in the loop" for decades. I think there might be some unexamined bias here. Certainly it's biased. I'm not the author, but to me there's a huge difference between computer/software as a tool, designed and planned, with known deterministic behavior/functionality, then put in the hands of humans, vs automating agency…

"Automating agency" it's such a good way to describe what's happening. In the context of your last paragraph, if they succeed in creating AGI, they won't be able to exercise control over a robot army, because the robot army will have as much agency as humans do. So they will have created the very situation they currently find themselves in. Sans an economy.

Re: I don't care how well your "AI" works

#267

Post by Aurornis »

bgwalter wrote: post #256
Aurornis wrote: post #234

Earlier quoted context omitted.

> I am actually less productive when using LLMs because now I have to read another entities code and be able to judge wether this fits my current business problem or not. You don’t have to let the LLM write code for you. They’re very useful as a smart search engine for your code base, a smart refactoring tool, a suggestion generator, and many other ways. I rarely have LLMs write code for me from scratch that I have t…

Is that why open source progress has generally slowed down since 2023? We keep hearing these promises, and reality shows the opposite.

> Is that why open source progress has generally slowed down since 2023?

Citation needed for a clam of that magnitude.

Re: I don't care how well your "AI" works

#268

Post by rsynnott »

catapart wrote: post #229
Spivak wrote: post #170

Earlier quoted context omitted.

I mean sure but none of these even claimed to help you do things you were already doing . If your job is writing code none of these help you do that. That being said the metaverse happened but it just wasn't the metaverse those weird cringy tech libertarians wanted it to be. Online spaces where people hang out are bigger than ever. Segways also happened they just changed form to electric scooters.

Being honest, I don't know what a 4GL is. But the rest of them absolutely DID claim to help me do things I was already doing. And, actually, NFTs and the Metaverse even specifically claimed to be able to help with coding in various different flavors. It was mostly superficial bullshit, but... that's kind of the whole tech for those two things. In any case, Segways promised to be a revolution to how people travel - so…

A 4GL is a "fourth generation language"; they were going to reduce the need for icky programmers back in the 70s. SQL is the only real survivor, assuming you're willing to accept that it counts at all. "This will make programmers obsolete" is kind of a recurrent form of magic tech; see 4GLs, 5GLs, the likes of Microsoft Access, the early noughties craze for drag-and-drop programming, 'no-code', and so forth. Even _COBOL_ was kind of originally marketed this way.

Re: I don't care how well your "AI" works

#270

Post by trashb »

In general using natural language to feed into AI to generate code to compile to runnable software seems like the long way around to designing a more usable programming language.

It seems that most people preferring natural language over programming languages don't want to learn the required programming language and ending up reinventing their own worse one.

There is a reason why we invented programming languages as an interface to instruct the machine and there is a reason why we don't use natural language.

Post reply on HN