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

Post-GPT Computing

grady.io

Post reply on HN
Print view View original

231–240 of 504 posts

Re: Post-GPT Computing

#231

Post by raincole »

raldi wrote: post #150
btbuildem wrote: post #142

Earlier quoted context omitted.

From your comment: > I don’t get the overall doom and gloom towards LLMs on the software field. From the second line of your comment: > Now you don’t need to hire junior devs Do you need GPT to put the two together? I think it's pretty obvious why folks are freaking out.

Now junior devs don't have to build the product of someone else's dreams; they can build the product of their own dreams.

... and starve? You know what kind of people create content of their own dreams? Artists. And the stereotype isn't "well-fed artists" for a good reason.

Re: Post-GPT Computing

#232

Post by spudlyo »

tarruda wrote: post #209

> Yesterday, I watched someone upload a video file to a chat app, ask a language model “Can you extract the first 5 s of the video?”, and then wait as the language model wrote a few lines of code and then actually executed that code Have we already solved AI safety problems? It seems like LLMs can now execute shell commands on our computers.

There is now a code interpreter[0] plugin for ChatGPT. It's not clear to me if it's available to folks who have been granted into the plugin alpha test or not, but it's running in a sandboxed execution environment somewhere -- not on our computers.

[0]: https://openai.com/blog/chatgpt-plugins#code-interpreter

Re: Post-GPT Computing

#233

Post by moomoo11 »

Are people asking it to generate code like “generate a random color hex code” or are they trying to use it to write code you’re going to put in production for users with access controls and various complexities?

Because yeah it works fine for basic programming things but I believe you need to know wtf you’re doing when it comes to anything more complex, even something basic like some of our single endpoint services.

Re: Post-GPT Computing

#234

Post by PeterisP »

tomp wrote: post #221
PeterisP wrote: post #216

Earlier quoted context omitted.

"they pretend to pay us, we pretend to work" is a failure because currently the prosperity of the society needs these people to actually work effectively. On the other hand, the hypothetical solution of "own the machines and distribute the wealth" is intended for a future which is substantially different, where it doesn't matter if everyone pretends to work or even explicitly avoids working, because that work isn't n…

Sure but then you don’t need wealth redistribution, because everything is dirt cheap. The correct mental model isn’t “communism” or “wealth distribution”, but instead “salt”. Countries used to go to war because of salt. But now technology has eroded its value so much, restaurants are literally giving it away.

Artificial scarcity is a thing, so something being dirt cheap to produce doesn't necessarily mean that it will be actually affordable, and in the current economic environment there seems to be sufficient motivation for powerful people to try and make various monopolies based on capital-gated barriers of entry, even if the marginal cost approaches zero; so I'd expect that the default scenario is not like "salt". Getting to a mental model "like salt" seems to be a reasonable outcome in the long run, but I'm afraid that it would take some significant pressure from the masses to get from here to there.

Re: Post-GPT Computing

#235

Post by crop_rotation »

tarruda wrote: post #209

> Yesterday, I watched someone upload a video file to a chat app, ask a language model “Can you extract the first 5 s of the video?”, and then wait as the language model wrote a few lines of code and then actually executed that code Have we already solved AI safety problems? It seems like LLMs can now execute shell commands on our computers.

They don't execute it on user's computers. They execute it on OpenAI computers.

Re: Post-GPT Computing

#236

Post by dento »

edgyquant wrote: post #48
meghan_rain wrote: post #8

Earlier quoted context omitted.

[flagged]

No it isn’t and this isn’t a funny joke. If you can’t write optimal algorithms and come up with data structures that allow for mapping out problems you aren’t solving anything and your job has always been one innovation away from disappearing.

I'd guess 90% dev jobs don't involve any algo knowledge more complicated than "should I pick a list or hashmap" or "which columns need indices". They involve converting business logic into code and combining it with good UI/UX.

Re: Post-GPT Computing

#237

Post by orangesite »

komposit wrote: post #181
angarg12 wrote: post #158

Last week I used ChatGPT for the first time for a real world task at work. It was a self-contained lambda function to perform some admin tasks, so it seemed like an ideal fit. Although the experience was good, it's far from the end of programmers. This was my experience: * Although ChatGPT is pretty good at generating code, it kept making simple mistakes such as calling non-existing APIs or introducing bugs. Some of…

This skepticism absolutely baffles me. Have you been using gpt-4? To unlock gpt for real you have to be careful to prompt it correctly and find a way to improve the feedback loop for improving code. It is only a matter of time until tools arrive that integrate this into your development environment and give it access to test/console output such that it can suggest code and iterate over the result. It's not perfect ye…

"You're holding it wrong"

Re: Post-GPT Computing

#238

Post by asdff »

rvz wrote: post #76
ape4 wrote: post #28

Earlier quoted context omitted.

And how can somebody become a senior dev unless they first worked as a junior

You become a 'senior dev" by running a company / startup yourself, since it is clear that after the tech layoffs, almost no-one has the money or profits to hire any new developers - junior and senior, which is why I say they are both affected. But even then, self-proclaimed seniors are too scared to start their own startup(s) now because of (1) Unfavourable market conditions (2) VCs hesitant to raise money (3) ChatGP…

I don't think starting a company will be that successful if you lack enough experience to be hired at any existing companies. The survivorship bias in tech is huge. Most things fail.

Re: Post-GPT Computing

#239

Post by golemotron »

yoyohello13 wrote: post #198
ticviking wrote: post #194

Earlier quoted context omitted.

I’m not terribly worried. Though I assume my work will begin to resemble that of the Adeptus Mechanicus rather than proper engineering. The fact is that anyone who understands even at a basic level what the computer is actually doing and isn’t afraid to look at it at a low level can’t be replaced by an AI trained on stack overflow. It may be that I will spend more of my time on code review of LLM generated code, or m…

> Though I assume my work will begin to resemble that of the Adeptus Mechanicus rather than proper engineering. Lol, I was thinking about this the other day. Eventually most devs will essentially just be praying to the Machine spirit to make the computer do what they want. A small few high clerics will bother to learn how computers actually work. The rest will simply be cargo culting to the maximum extent possible.

> Eventually most devs will essentially just be praying to the Machine spirit to make the computer do what they want.

Same as it ever was.

Re: Post-GPT Computing

#240

Post by raincole »

piokoch wrote: post #74

Ok, so we have that software written by AI, AI is clever, it does not need good variable names or functions/methods/class names, some of the stuff it will call according to passed specification so it will be understandable, but the further it goes, everything will get more generic, taken from kirjillion of other code snippets on Github. And it all will be working. Until someone starts testing this and finds a bug. An…

I believe you never use ChatGPT. I'm not claiming you never use GPT4, I'm claiming you never use even GPT3.5. If you did you will notice its problem is the opposite of what you describe. Especially:

> it does not need good variable names or functions/methods/class names,

It's the exact opposite. It's too good at naming things. It insists to use variable/function names that make sense in plain english, and often make mistakes when the API has inconsistent naming, or consistent but unusual naming.

For example, it makes mistakes when writing code that use "Loop" in Blender API. And the reason is quite obvious to me: because Blender's "Loop" is not what loop means in plain english.

Post reply on HN