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

I wish GPT4 had never happened

chaudhry.notion.site

Post reply on HN
Print view View original

531–540 of 1001 posts

Re: I wish GPT4 had never happened

#531

Post by bmn__ »

kofejnik wrote: post #473
HyperSane wrote: post #450

The fact that so many people fear loss of income as a result of better AI is a real indictment of our economic system. A better system that more fairly distributed economic productivity would make people welcome better AI because they could be confident it would make everyone wealthier instead of just a few people rich enough to be Goa'uld.

Define “fairly”

The video "Wealth Inequality in America" https://youtu.be/QPKKQnijnsM?t=35> at timestamp 0:35 shows a definition most respondents to the question agree on.

Re: I wish GPT4 had never happened

#532

Post by renewiltord »

pavlov wrote: post #492
kofejnik wrote: post #473

Earlier quoted context omitted.

Define “fairly”

The grandparent comment does contain a definition: “A […] system that […] would make people welcome better AI because they could be confident it would make everyone wealthier” That would be a fair system regardless of the specifics of how it’s achieved.

Well, then we have been living in a system that's as close to fair as possible since global individual wealth has been growing for a long time. The few times we had it shrink significantly over some period of time is when local populations attempted forced wealth distribution.

Our present system, by this definition, is amazing!

Re: I wish GPT4 had never happened

#533

Post by passion__desire »

manmal wrote: post #514
throw__away7391 wrote: post #457

Earlier quoted context omitted.

Shell scripting and JavaScript are strengths, any kind of server-side language like Scala or C# it really falls apart quite quickly. I think that integrating into existing APIs is the Achilles' heel, GPT-4 doesn't seem to do any better here either. It makes me wonder if someday we might not get better results completely eschewing libraries altogether and letting the transformers generate literally everything--perhaps…

I think it will be quite the opposite. Generating a lot of code means generating a lot of bugs. LMMs should get the (battle tested) libs they have to use as part of the prompt, and they should just use those APIs.

Just as nobody should roll their own cryptographic library, the same rules apply to LLMs.

Re: I wish GPT4 had never happened

#534

Post by nomel »

nlh wrote: post #250
surgical_fire wrote: post #156

Funny thing - I've been using ChatGPT as a programming assistant. My experience with it is that the more knowledgeable I am at the domain, the more useful it is. It often generates code that won't compile or won't run correctly, but that sort of goes in the correct direction, enough that I can spot the blunders and refine my questions. Quite often I even learn a few new things when using it to explore libraries and f…

I've had a similar experience (but slightly different). The biggest innovation that ChatGPT has brought to me in my programming/workflow tasks is the comfort and ease in which it creates shell scripts. That's an area that's always been just past my comfort zone ("Create a shell script to solve this problem? Now I have two problems..."). But all of a sudden, I feel like a master at shell scripting! Things that have al…

Using it with “old” systems, like bash and objective-c, is why I decided to shell out $20/month. It easily saves me $1.25 of pain, per work day.

But, it also really highlights that it’s trained on data, and doesn’t “know”. I’ve never seen it produce a bash script/line that properly handled spaces, on its own, because most shell scripts don’t. Even ChatGpt 4.0 backed fails to “correct” the mistakes, a large portion of the time.

Re: I wish GPT4 had never happened

#535

Post by saulpw »

carapace wrote: post #412
__MatrixMan__ wrote: post #251

Earlier quoted context omitted.

> This has been the promise over and over again And over and over again, we've underestimated and then normalized the capacity of our institutions, made of otherwise decent people, to achieve outcomes that we would call abhorrent if they were carried out by an individual. So the door closes on that opportunity because we fail to act. If we're ever going to walk through that door, companies need to fear us, because th…

> they sure as hell aren't going to share with us out of the goodness of their hearts You know you can buy stock, right? And, as a shareholder, those companies will give you the profit?

Where do I go to buy stock in OpenAI?

Re: I wish GPT4 had never happened

#536

Post by grumple »

School-Cotton wrote: post #310
rootusrootus wrote: post #287

I've been using ChatGPT a lot. I'm no longer feeling much anxiety over the future of software engineering. GPT is a great tool. Still just a tool. It helps remind me how to do things I do rarely. To get code that exactly fits my requirements, I'd have to get so specific that I might as well just write the silly code myself. But for shell scripts, SQL, stuff like that, it's pretty decent. Yesterday I asked it to give…

Are you using 3.5 or 4? If 3.5, your experience is irrelevant since 4 is so much better.

I haven't noticed much of a difference. Can you give some examples where 4 has been much better?

Re: I wish GPT4 had never happened

#538

Post by vc8f6vVV »

pydry wrote: post #283
mort96 wrote: post #249

Earlier quoted context omitted.

> This has been the promise over and over again, for centuries, and it has consistently not paid off. Where's the predicted society where automation allows us all to work for two hours a day, and spend the rest at leisure? The extra value has gone to enrich the oligarchs rather than to the people. Maybe that should change.

How do you deprive an oligarch of their ill gotten wealth without violence?

Think one step further. What happens if you do? Doesn't the USSR example teach you anything? You will have a different breed of oligarchs, that's all (yes, there were oligarchs in the USSR, just not in the way you think, money is not everything and oligarchy is not about money).

Re: I wish GPT4 had never happened

#539

Post by Aaargh20318 »

phkahler wrote: post #505
Aaargh20318 wrote: post #29

The big oversight here is that Priya does not want a job , what Priya wants is an income . The point of life is not working, the point of life is to enjoy it. We should strive to eliminate as many jobs as we possibly can, the less we have to work the better it is. To make this possible we need to find a way to a new system that doesn't directly link labour with income .

>> To make this possible we need to find a way to a new system that doesn't directly link labour with income. What is income? What is it used for? Answers: it's a reward for work. And it is used as an incentive to get others to work.

> What is income? What is it used for?

Money is a proxy for resources. If you boil it down to the essentials, income is the way we allocate the resources available to us to different individuals. It is not intrinsically a reward for work. It isn't even directly linked to work anymore in our capitalist system. The people who work the hardest don't have the highest income. The way the system currently works is that those with the most resources get the most additional resources, simply because they already have a lot of them.

Re: I wish GPT4 had never happened

#540

Post by mnky9800n »

surgical_fire wrote: post #156

Funny thing - I've been using ChatGPT as a programming assistant. My experience with it is that the more knowledgeable I am at the domain, the more useful it is. It often generates code that won't compile or won't run correctly, but that sort of goes in the correct direction, enough that I can spot the blunders and refine my questions. Quite often I even learn a few new things when using it to explore libraries and f…

This exactly. I've been using it to write R code because I've never learned R and needed to quickly write a lot of it. This kind of worked because when I get stuck on a bug that I can't figure out that was generated by chatgpt then my colleague helps me. But really this leaves me with questions like, what if it didn't make so many bugs? What is there was an interface that just evaluated the code I asked it to write? What if most code writing jobs become prompt engineering jobs? I agree the more expert I am the better expert chatgpt seems to be. It will never know what question I am thinking. But it does seem to accelerate or replace a lot of human effort.
Post reply on HN