Live data from Hacker News

I wish GPT4 had never happened

chaudhry.notion.site

481–490 of 1001 posts

Re: I wish GPT4 had never happened

#481

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 .

What a profoundly naive opinion. Most people find meaning, fulfillment, even contentment in working, especially when the problems are engaging and the solutions produce value. Many people claim their lives would be better if they never had to work, but they are bullshitting themselves with childhood naivete. Not working is a sucking void of boredom, nihilism, hedonism, and despair. I've been through some of that befo…

I don’t know, I think maybe the midwit meme applies here.

(Early) retirement is an adjustment. It takes time to find them, but there are alternative ways of finding fulfillment and I’m happy to have them. N=1 but my revealed preference seems to be that yes, being financially secure is good, it reduces working, and it’s possible to find better alternatives.

Re: I wish GPT4 had never happened

#482

Earlier quoted context omitted.

"Work" above is being used in a more specific form than you are using it. GP is not referring to all productive work, they are specifically referring to employment on the market. Yes, human beings are better in relationships and better when they help each other, but GP is not talking about abolishing helping other people or doing meaningful things. There are a huge number of things people can do to meaningfully make…

I your work is not producing anything valuable for other people it's void and you will get bored pretty fast, it's a human nature. If it does you can always find a way to monetize it as people need it and ready to share their wealth for something they need. No, it doesn't have to be a work "for a company", but if you can't "make money off of this" it probably means your "work" is only relevant to you, or somebody els…

> I your work is not producing anything valuable for other people it's void and you will get bored pretty fast, it's a human nature

I think this is plain wrong. There are plenty of pursuits that are “work” on the mental or physical side, but do not produce economic value. Developing artistry by mastering a medium or an instrument, climbing mountains, running ultra marathons, tending to a garden or bonsai, modifying a vehicle… These are things could produce economic or social credit, but are largely solo “work” which can be (and often is) self satisfying without external motivation.

Re: I wish GPT4 had never happened

#483

I asked GPT4 to write me a simple document. Took like 9 tries and I had to proof read it and make adjustments. And that's with a simple paragraph in a written language. Forget about doing the header and the footer properly. It's silly that I have to spell it out on HN of all places. Writing code is much more complex. There's a reason we call the job "Software Engineer". GPT doesn't know what it doesn't know. It doesn…

Writing code is actually much more easy. Since there is often clear rules of how to do things. Writing a good text is much harder. GPT is great at writing code, terrible at writing a proper text. Also might have to do with the effort OpenAI is putting in reinforcing the code quality with Senior devs. It's writing might also get better if it would solely use senior writers to give feedback.

Re: I wish GPT4 had never happened

#484

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…

> My experience with it is that the more knowledgeable I am at the domain, the more useful it is.

Exactly right. I made this argument in an unrelated thread:

I gave an example of testing ChatGPT by asking it to write CRC calculation code in ARM assembler. The code worked. It wasn’t as good as mine (slower, inefficient), yet, as you said, as a starting point, it would have been fine.

“This is today, mid 2023. Give it a year or two (maybe less?) and it will be a tool to contend with. People who like to blame everything else rather than their lack of knowledge and experience will not do very well in that world.

Why would I pay someone to do when they bring nothing special to the table?

Here's the huge paradigm shift (at least for me):

I could not care less what someone knows or does not know. I care about the range and breath of their experience and how they approach learning that which you do not know.

Someone like that can use any available tool, including AI tools, to deliver value in almost any domain. Someone who blames others (tools, people, the system, whatever), cannot.

We might just be entering an era in which experience will be shown to have serious value.”

Re: I wish GPT4 had never happened

#485
post #227

I asked GPT4 to write me a simple document. Took like 9 tries and I had to proof read it and make adjustments. And that's with a simple paragraph in a written language. Forget about doing the header and the footer properly. It's silly that I have to spell it out on HN of all places. Writing code is much more complex. There's a reason we call the job "Software Engineer". GPT doesn't know what it doesn't know. It doesn…

When we built the first airplane it could barely carry a single person and didn't go far... We are just barely touching digital general intelligence, and we're not even using API enabled versions that use tooling. And that's what we have today. Tell me, what is AI going to do tomorrow? >If you're worried about GPT taking over your job you just show your own lack of knowledge and experience. Oh, I'm glad you popped ou…

> When we built the first airplane it could barely carry a single person and didn't go far... We are just barely touching digital general intelligence, and we're not even using API enabled versions that use tooling.

Big assumption that we're not just going to slam into a wall of capabilities and end up just with some really good autocomplete machines.

(I am bullish on generative AI as a whole, but come on, you have no clue about what's coming next just like me)

Re: I wish GPT4 had never happened

#486
post #348

Earlier quoted context omitted.

>I see LLM models not as the end of society, This is because you're used to normal technological development scaling. Unfortunately with AI we are unsure at this point if our old paradigms still apply. What happens with society really depends on where we are on the technology growth curve. If LLMs plateau for a while and we don't see significant growth in their abilities, then we'll just see pretty massive technologi…

That's because you're used to normal technological development hype. Every few years a shiny new toy enters the hype cycle, only to fall out of favor as things stop advancing as people expected. LLMs are impressive at natural language processing (which is a damn hard problem to solve), and can both parse and generate natural language at a satisfactory level. It's an interesting facet of AI, but it also has some glari…

It's almost as if you didn't parse most of my reply.

Again, if we are close to AGI every single one of your past ideas on how technology should behave is outdated and done for.

I didn't say we were going to get AGI, and I said it's completely possible we are going plateau out. But if we create general intelligence it will be as if we have created fire, or the wheel, it will present a profound and drastic change in technological capabilities that it will reshape all of humanity.

Re: I wish GPT4 had never happened

#487

If you don't like these fast changes, time to plan a move to more "traditional" countries like Italy or Spain. Apart from fiber internet, almost all tech advances were ignored for the most part: limited uber, limited airbnb, and yes you will need to call the restaurant to book a table. Even just the concept of "chat support" is foreign, let alone having a website for a business. There are still very many businesses w…

This really is the way. What use are politicians if they don't make laws that make people's lives easier. To bad the United States isn't a democracy like many European countries are.

How does blocking Uber make people's lives easier? Or did you just mean easier for the taxi drivers because they can extract political rents using the government as their tool, at the expense of broader society?

Re: I wish GPT4 had never happened

#488

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 .

>> 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.

As humans we have not found a way to allow people not to work. Anyone not working is either living off money from their own prior work or benefitting from the work of someone else. Let's see and validate such a system before we run off eliminating jobs.

Re: I wish GPT4 had never happened

#489
Genuine question, how does one get over the 30k token limit in a gpt4 request in such cases?

From my personal experience, it seems while today we can create PoCs of GPT4 doing a white colar job replacement, when you then try to actually productize it you still hit a data encoding limit...

And that doesn't even touch the QoS topics that come with automation/ml.

Re: I wish GPT4 had never happened

#490

1) Chaudry's article is amazing in its lack of creative thinking. If people like Priya can be ~10 times more productive wielding some AI tools, that means that many companies can potentially produce 10 times as many products/services and make more things that will enrich peoples' lives. Though some people in some companies might be laid off, it doesn't necessarily mean that more people will be laid off than new jobs…

> that means that many companies can potentially produce 10 times as many products/services and make more things that will enrich peoples' lives. 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? > Though some people in some companies might be laid…

> 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

That's the opposite of what GP predicted: that people would be more productive and produce more, not work less.

Programmers are way more productive than the were sixty years ago. We don't have fewer programmers, or part-time programmers, but we do have vastly more and better software.

Post reply on HN