We are beginning to roll out new voice and image capabilities in ChatGPT • Hacker News Board

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#711

Post by mvdtnz » Mon, Sep 25, 2023, 8:09 PM UTC

callwhendone wrote: post #532

RivieraKid wrote: post #466

I went from being worried to thinking it won't replace me anytime soon after using GPT4 for a while and now I'm back to being worried. Because the pace of development is intense. I would love to be financially independent and watch this with excitement and perhaps take on risky and fun projects. Now I'm thinking - how do I double or triple my income so that I reach financial independence in 3 years instead of 10 year…

I'm very worried constantly. This is the story of the bear, where you just have to be faster than the other guy. For now. The bear is getting faster and faster and it won't be long before it eats all of us. It feels like we're at the end of history. I don't know where we go from here but what are we useful for once this thing is stuck inside a robot like what Tesla is building? What is the point of humanity? Even tak…

I think you need to calm down. "AI" today is just fancy predictive text and there's no evidence it will progress far beyond that.

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#712

Post by Karrot_Kream » Mon, Sep 25, 2023, 8:10 PM UTC

WendyTheWillow wrote: post #692

famouswaffles wrote: post #681

Earlier quoted context omitted.

>the gap between 3 and 4 is not anything as large as the gap between 2 and 3. We'll just have to agree to disagree. 3 was a signal of things to come but it was ultimately a bit of a toy, a research curiosity. Utility wise, they are worlds apart. >if there really was no innovation at all as you claim, then there will be a limit hit at computing capability and cost. computing capability and cost are just about the one…

Do you realize I'm not disagreeing with you about the difference between 3 and 4? Reread what I wrote. I contrasted 3 and 4 with 2 and 3, which you seem to be entirely ignoring. 3 and 4 could be worlds apart, but wouldn't matter if 2 and 3 were two worlds apart, for example. And it is not true that computing power will continue to reduce; Moore's Law has been dead for some time now, and if incremental growth in LLMs…

> And it is not true that computing power will continue to reduce; Moore's Law has been dead for some time now, and if incremental growth in LLMs require exponential growth in computing power the marginal difference won't matter.

I think the stronger argument here won't necessarily be Moore's Law related but a change in architecture. Things like Apple's Neural Engine, Google's TPMs, or Geohot's Tinybox. In Intel's Tick-Tock model, this is the Tock for the previous Tick of larger datasets so to speak.

(Note: I don't necessarily agree, just trying to make a stronger argument than just invoking Moore's Law.)

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#713

Post by jabradoodle » Mon, Sep 25, 2023, 8:12 PM UTC

TheRoque wrote: post #596

jader201 wrote: post #565

Earlier quoted context omitted.

This feels fairly naive, ignoring how much progress has happened over the (short) span of one year. This doesn't sound like that tough of a gap to close in another year (again, projecting based off recent progress).

The opposite is fairly naive. Software development is not only dumping tokens into a text file. To have a significant impact on the market, it should do much, much, much more: compile and test code, automatically assess the quality of what its done, be aware of the current design trends (if in UI/UX), ideally innovate, it should also be able to run a debugger, inspect all the variables, and deduce from there how it g…

In a market you don't need to replace it, you replace millions of hours of building/testing/documenting apps and you've effectively reduced demand for SWE labour (ignoring the potential for new job creation).

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#714

Post by brvsft » Mon, Sep 25, 2023, 8:12 PM UTC

LambdaComplex wrote: post #529

make3 wrote: post #521

Earlier quoted context omitted.

The real problem is distribution of the output of production. We will need something like UBI eventually.

UBI is a bandaid on top of capitalism. It is saying "we have a system where people die if they don't have money, so we'll give people money." It's not a real fix. A real fix would be replacing the system with one where people don't need money in order to not die. We're going to keep automating more and more things. I think that much is inevitable. Eventually, we may get to a point where very few jobs are necessary fo…

People don't need money to not die, they need to work to not die. Until the fulfillment of everyone's basic needs can be automated, people are expected to work (disregarding some obvious exceptions). The money is just a proxy for that work.

Although there is certainly a lot of fuckery going on with the money (currency) itself, but if that's the problem you're alluding to, I don't think summarizing it as "capitalism" is accurate.

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#715

Post by kortex » Mon, Sep 25, 2023, 8:13 PM UTC

pmarreck wrote: post #661

mhss wrote: post #604

Earlier quoted context omitted.

People that has enough money & wealth to employ others, control corporations who employ thousands and influence the government meaningfully (lobbying). Today, they basically control who is employed and who isn't by deciding how much to invest and in what. No, you cannot just choose to employ yourself in all circumstances. I'm all for being self-sufficient but not everyone can start a company (time and capital intensi…

If you don’t believe you can do a thing before you even try, then you’re correct, right out of the gate. (insert eyeroll emoji here) All your repeated uses of “they” points to a toxic external-locus-of-control worldview. You were always the only limit of yourself. Any other claim amounts to heretical self-dehumanization. You’re not fungible and never were, and anyone who tries to make you believe that deserves the ut…

> You’re not fungible and never were

Sure as heck doesn't feel that way. And that's as a software developer with multiple college degrees and a decade of experience. The neurodivergence means I've always had to mask and be on guard and push well beyond my limits into physical/mental damage, because the fear of losing employment is ever-present. Feels pretty commoditized.

> and anyone who tries to make you believe that deserves the utmost vehement pushback.

The faceless corporations and their boards of investors who value "line go up" over basically every other metric of human wellbeing? Yes, they absolutely deserve pushback, but it's not easy (open source COTS guillotine plans, anyone?).

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#716

Post by ryandrake » Mon, Sep 25, 2023, 8:13 PM UTC

whatamidoingyo wrote: post #683

callwhendone wrote: post #673

Earlier quoted context omitted.

I don't know what your salary is but mine isn't going to be replaced by demoing a bathroom and I have a mortgage and a standard of living I was hoping to be able to afford at least until my kids are out of the house.

Unless you're making a ridiculous amount of money, you can definitely match a developer salary remodeling homes. So long as you're the actual business owner. This was just an example, of course.

But being the actual business owner is also not "demolishing a bathroom and installing tile." While AI won't be able to physically demo a bathroom, it's conceivable that it will eventually be able to run a business better than a human entrepreneur.

The only jobs that seem to be safe (for the medium term) are jobs that require some physical manipulation of the world. Like, the actual, hands-on physical work that tradespeople do. Although they'll eventually fall to AI-powered robots.

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#717

Post by toomuchtodo » Mon, Sep 25, 2023, 8:17 PM UTC

idopmstuff wrote: post #244

suyash wrote: post #29

This announcement seem to have killed so many startups that were trying to do multi-modal on top of ChatGPT. The way it's progressing with solving use cases with images and voice, not too far when it might be the 'one app to rule them all'. I can already see "Alexa/Siri/Google Home" replacement, "Google Image Search" replacement, ed-tech startups that were solving problems with AI using by taking a photo are also doo…

It increasingly feels to me like building any kind of general-use AI tool or app is a bad choice. I see two viable AI business models: 1. Domain-specific AI - Training an AI model on highly technical and specific topics that general-purpose AI models don't excel at. 2. Integration - If you're going to build on an existing AI model, don't focus on adding more capabilities. Instead, focus on integrating it into compani…

If you focus on integration, you're up against autogpt, gorilla, etc.

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#718

Post by pmarreck » Mon, Sep 25, 2023, 8:17 PM UTC

kortex wrote: post #705

pmarreck wrote: post #663

Earlier quoted context omitted.

That’s not going to come to fruition, and no amount of dreamy socialist fanfiction’ing is going to make it so. People pay for value. Produce value for others, get paid. LLM’s are tools to make humans able to produce more value, and will not replace humans, although the job market will change, and hopefully utilize humans better. People, NOT machines, are the ultimate judgers of what is valuable and the ultimate produ…

> Produce value for others, get paid So if a human is unable to produce value, they don't get (food/education/heathcare/ )? That seems to be the implication. We in developed countries already have some amount of "value risk hedging" (I'm loathe to say "socialism" here), we just disagree endlessly how much is the optimal amount. But we've determined that wards of the state, universal education, and some amount of food…

> So if a human is unable to produce value, they don't get (food/education/heathcare/)?

I would agree that some people are simply unable to help and need the help themselves and should get it. UBI or some other social safety net should be there for that.

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#719

Post by jabradoodle » Mon, Sep 25, 2023, 8:17 PM UTC

doug_durham wrote: post #484

NikolaNovak wrote: post #291

Earlier quoted context omitted.

This is a fairly perpetual discussion, but I'll go for another round: I feel like using LLM today is like using search 15 years ago - you get a feel for getting results you want. I'd never use chatGPT for anything that's even remotely obscure, controversial, or niche. But through all my double-checking, I've had phenomenal success rate in getting useful, readable, valid responses to well-covered / documented topics s…

Exactly this! This is my experience also. Your point about "well covered & non-controversial" is spot on. I know not to expect great results when asking about topics that have very little coverage. To be honest I wouldn't expect to go to an arbitrary human and get solid answers on a little covered topic, unless that person just happened to be topic expert. There is so much value in having the basics to intermediate l…

I think a key difference is that humans very rarely sound convincing talking about subjects they have no clue about.

I've seen the hallucination rate of LLMs improve significantly, if you stick to well covered topics they probably do quite well. The issue is they often have no tells when making things up.

Re: We are beginning to roll out new voice and image capabilities in ChatGPT

#720

Post by charcircuit » Mon, Sep 25, 2023, 8:18 PM UTC

whimsicalism wrote: post #583

charcircuit wrote: post #506

Earlier quoted context omitted.

>From my firm conviction 18 months ago that this type of stuff is 20+ years away; It was totally possible. There just was not a consumer facing product offering the capability.

I disagree that at current possibility it was "totally possible" but it was 100% obvious by that point that it was going to be possible very soon . IMO that has been clear since ~2019.

GPT3 existed. OCR existed. Object recognition existed.