zlacker

[parent] [thread] 34 comments
1. teeker+(OP)[view] [source] 2023-03-01 15:55:27
Really? I feel like they'll go the way of Docker, but faster: Right now super hot, nice tools/API, great PR. But it's build on open and known foundations, soon GPTs will be commodity and then something easier/better FOSS will arise. It may take some time (2-3 years?) but this scenario seems most likely to me.

Edit: Ah didn't get the "reference", perhaps indeed it will be the last of the tech companies ever indeed, at least one started by humans ;).

replies(2): >>startu+md >>jrm4+jm
2. startu+md[view] [source] 2023-03-01 16:48:00
>>teeker+(OP)
Possible. Coding as we know it might get obsolete. And it is a trillion dollar industry.
replies(3): >>teeker+xj >>fsckbo+lX >>mckrav+Cw2
◧◩
3. teeker+xj[view] [source] [discussion] 2023-03-01 17:09:51
>>startu+md
I have ChatGPT make up some code every now and then. It’s really nice and when not obscure usually directly useable. By you need to understand what it produces imo. I love that it also explains the code and I can follow code it generates and judge its quality and applicability. Isn’t that important?
replies(1): >>Fillig+sn
4. jrm4+jm[view] [source] 2023-03-01 17:19:44
>>teeker+(OP)
Are there any companies like Docker? They feel like such a Black Swan to me; namely -- "billions" + very useful + pretty much "not evil" at all. I literally can't think of any others.

I don't mean "literally evil" of course, I personally acknowledge the need to make money et al, but I mean it even seems like your most Stallman-esque types wouldn't have too much of a problem with Docker?

replies(3): >>throwa+6F >>ynx+qL >>goodpo+Dh1
◧◩◪
5. Fillig+sn[view] [source] [discussion] 2023-03-01 17:24:02
>>teeker+xj
Last year the output was poor. The year before then, GPT essentially couldn't write code at all...

You're not wrong about its quality right now, but let's look at the slope as well.

replies(2): >>buddhi+Lr >>jay_ky+dM
◧◩◪◨
6. buddhi+Lr[view] [source] [discussion] 2023-03-01 17:38:29
>>Fillig+sn
on the other hand, gpt-3 was trained on a data set that contains all of the internet already. A big Limitation seems to be that it can only work with problems that it has already seen
replies(2): >>pixl97+GG >>JohnFe+qI
◧◩
7. throwa+6F[view] [source] [discussion] 2023-03-01 18:24:36
>>jrm4+jm
Netflix?
replies(2): >>jrm4+NX >>majou+rg2
◧◩◪◨⬒
8. pixl97+GG[view] [source] [discussion] 2023-03-01 18:29:30
>>buddhi+Lr
I mean a huge amount of code I see is stuff like "get something from API, do this, and pass to API/SQL" so I'm assuming a lot of that could be automated.
replies(1): >>ethbr0+cO1
◧◩◪◨⬒
9. JohnFe+qI[view] [source] [discussion] 2023-03-01 18:36:17
>>buddhi+Lr
> on the other hand, gpt-3 was trained on a data set that contains all of the internet already

This fact has made me cease to put new things on the internet entirely. I want no part of this, and while my "contribution" is infinitesimally small, it's more than I want to contribute.

◧◩
10. ynx+qL[view] [source] [discussion] 2023-03-01 18:49:48
>>jrm4+jm
While I have no problem with Docker, it probably is worth noting that their entire product is based on a decade and a half of Google engineering invested into the Linux kernel (cgroups).

Not that that demeans or devalues Docker, but it contextualizes its existence as an offshoot of a Google project that aimed to make it possible (and did, but only internally)

replies(1): >>jankey+I51
◧◩◪◨
11. jay_ky+dM[view] [source] [discussion] 2023-03-01 18:52:26
>>Fillig+sn
The whole model is wrong. We don't need an AI that just spits out words that look like words its seen before. We need it to understand what its doing.

Midjourny needs to understand that its drawing a hand, that hands have 5 fingers.

Nobody will use Bing chat for anything important. Its like asking some some random guy on the train. He might know the answer, and if its not important then fine, but if it is important, say the success of your business, you're going to want talk to somebody who actually knows the answer.

replies(1): >>rvnx+Tr1
◧◩
12. fsckbo+lX[view] [source] [discussion] 2023-03-01 19:51:53
>>startu+md
> Coding as we know it might get obsolete. And it is a trillion dollar industry.

freeing up that many knowledge workers to do other things will grow the economy, not shrink it, a new industrial revolution

replies(3): >>rnk+z11 >>elzbar+e71 >>mitthr+4g1
◧◩◪
13. jrm4+NX[view] [source] [discussion] 2023-03-01 19:54:29
>>throwa+6F
Don't know why this was downvoted, I tend to agree. There's a clear relationship between the money and the product. Pay money, receive access to a bunch of shows. (as opposed to, e.g. SEO click-view advertising shadiness)
◧◩◪
14. rnk+z11[view] [source] [discussion] 2023-03-01 20:12:30
>>fsckbo+lX
That's an excellent point that I don't think is made enough. The great pay and relative freedom software engineering provides the technically-minded people of the world is great for us, yet starves many other important fields from more technical innovation because of not enough workers in those fields.
replies(1): >>robert+wa1
◧◩◪
15. jankey+I51[view] [source] [discussion] 2023-03-01 20:32:56
>>ynx+qL
So is OpenAI actually, GPT -> transformer, invented at Google. DALL-E -> diffusion, invented at ... Google.
replies(1): >>margor+Ym1
◧◩◪
16. elzbar+e71[view] [source] [discussion] 2023-03-01 20:39:51
>>fsckbo+lX
Exactly what?
◧◩◪◨
17. robert+wa1[view] [source] [discussion] 2023-03-01 20:54:18
>>rnk+z11
Can you explain? I think I'm reading it wrong, but it seems as though you're saying the presence of something is what causes starvation of that thing.
replies(1): >>fsckbo+Nm1
◧◩◪
18. mitthr+4g1[view] [source] [discussion] 2023-03-01 21:22:57
>>fsckbo+lX
Yes; hopefully they get good at working with their hands.
replies(1): >>Camper+Ij1
◧◩
19. goodpo+Dh1[view] [source] [discussion] 2023-03-01 21:32:10
>>jrm4+jm
> Are there any companies like Docker?

Let's hope not.

> "not evil" at all

Sarcasm?

replies(2): >>DiggyJ+Lq1 >>jrm4+i42
◧◩◪◨
20. Camper+Ij1[view] [source] [discussion] 2023-03-01 21:43:58
>>mitthr+4g1
Well, yes, because then maybe I won't have to pay $250/hour to have my car fixed.
◧◩◪◨⬒
21. fsckbo+Nm1[view] [source] [discussion] 2023-03-01 22:00:58
>>robert+wa1
no, he's saying other areas that could use smart people are being starved of a certain type of smart, analytical people, for example the types or quantities of people who might be attracted to the medical field.

Lester Thurow, a pretty left liberal economist, pointed out that women's "liberation" and entrance into the general workforce had starved teaching/education of the pool of talented women that had previously kept the quality of education very high. (His mother had been a teacher.)

I (who had studied econ so I tend to think about it) noticed at the dawn of the dot-com boom how much of industry is completely discretionary even though it seems serious and important. Whatever we were all doing before, it got dropped so we could all rush into the internet. The software industry, which was not small, suddenly changed its focus, all previous projects dropped, because those projects were suddenly starved of capital, workers, and attention.

replies(2): >>rnk+kl5 >>robert+uui
◧◩◪◨
22. margor+Ym1[view] [source] [discussion] 2023-03-01 22:01:57
>>jankey+I51
Seems like maybe many of their products are mediocre and short-lived but their engineering and research are top-notch.
◧◩◪
23. DiggyJ+Lq1[view] [source] [discussion] 2023-03-01 22:26:29
>>goodpo+Dh1
What's evil about Docker (the company)?
replies(1): >>matt_h+Fz1
◧◩◪◨⬒
24. rvnx+Tr1[view] [source] [discussion] 2023-03-01 22:33:18
>>jay_ky+dM
The new models similar to Midjourney (Stable diffusion) and notably Deliberate v3 can perfectly draw anatomy now and you can even choose how many fingers you want.
◧◩◪◨
25. matt_h+Fz1[view] [source] [discussion] 2023-03-01 23:20:18
>>DiggyJ+Lq1
https://news.ycombinator.com/item?id=28369570

https://news.ycombinator.com/item?id=27013865

replies(2): >>goodpo+cC1 >>DiggyJ+Eo4
◧◩◪◨⬒
26. goodpo+cC1[view] [source] [discussion] 2023-03-01 23:35:03
>>matt_h+Fz1
The bait and switch was not nice but there's more...
◧◩◪◨⬒⬓
27. ethbr0+cO1[view] [source] [discussion] 2023-03-02 01:09:41
>>pixl97+GG
OrmGPT!
◧◩◪
28. jrm4+i42[view] [source] [discussion] 2023-03-02 03:36:30
>>goodpo+Dh1
Not awesome, but nowhere near the definite evil I've seen from Microsoft, Oracle, Amazon, etc....
◧◩◪
29. majou+rg2[view] [source] [discussion] 2023-03-02 05:40:32
>>throwa+6F
The CEO declared sleep their competitor.
◧◩
30. mckrav+Cw2[view] [source] [discussion] 2023-03-02 08:29:28
>>startu+md
I was initially impressed and blown away when it could output code and fix mistakes. But the first time I tried to use it for actual work, I fed it some simple stuff and I even had the pseudo code as comments already - all it had to do is to implement it. It made tons of mistakes and trying to correct it felt like way more effort than just implementing it myself. Then that piece of code got much more complex, and I think there's no way this thing is even close to outputting something like that, unless it has seen it already. And that was ChatGPT, I have observed Copilot to be even worse.

Yes, I'm aware though it may get better but we actually don't know that yet. What if it's way harder to go from outputting junior level code with tons of mistakes to error-free complex code, than it is to go from no capability to write code to junior level code with tons of mistakes? What if it's the difference between word prediction algorithm and actual human-level intelligence?

There may be a big decrease in demand, because a lot of apps are quite simple. A lot of software out there are "template apps", stuff that can theoretically be produced by a low code app, will be eventually produced by a low code app, AI or not. When it comes to novel and complex things, I think it's not unreasonable to consider that the next 10 - 20 years will still see plenty demand for good developers.

replies(1): >>startu+mL3
◧◩◪
31. startu+mL3[view] [source] [discussion] 2023-03-02 16:47:56
>>mckrav+Cw2
Considering that OpenAI started instruction following alignment a month ago, with 1k workers, to do engineering tasks, coding might be solved now.
replies(1): >>mckrav+b74
◧◩◪◨
32. mckrav+b74[view] [source] [discussion] 2023-03-02 18:17:40
>>startu+mL3
I have just decided to give it another try on a very straightforward thing. I have asked it to get me a Rust function that uses a headless browser to get the HTML of a fully loaded webpage.

ChatGPT:

let screenshot = tab.capture_screenshot(ScreenshotFormat::PNG, None, true).await?;

let html = String::from_utf8(screenshot).unwrap();

>[...] Once the page is fully loaded, the function captures a screenshot of the page using tab.capture_screenshot(), converts the screenshot to a string using String::from_utf8(), and then returns the string as the HTML content of the page.

Of course, it admits to the mistake (sort of, it still does not get):

> You are correct, taking a screenshot and converting it to text may not always be the best approach to extract the HTML content from a webpage.

It's hilarious.

◧◩◪◨⬒
33. DiggyJ+Eo4[view] [source] [discussion] 2023-03-02 19:29:02
>>matt_h+Fz1
That is very far away from my personal definition of the word "evil".
◧◩◪◨⬒⬓
34. rnk+kl5[view] [source] [discussion] 2023-03-03 00:45:11
>>fsckbo+Nm1
Yes, you are correct, that's what I was trying to say.
◧◩◪◨⬒⬓
35. robert+uui[view] [source] [discussion] 2023-03-07 12:56:14
>>fsckbo+Nm1
Ah, I see. Thanks.

In terms of doctors, I think there is a counterbalancing effect of sorts, whereby some administration can be digitised and communication is more efficient, but it probably doesn't make up for the lack of additional candidates.

[go to top]