It’s been incredibly annoying seeing so many new grifters pop up seemingly out of nowhere as “AI experts” and the like, promising (and even guaranteeing) random folks that they’ll make millions jumping on this bandwagon. It reminds me of how insufferable so much of the get-rich-quick crypto content was in 2017 and then in 2020/2021 (even as a crypto believer myself).
Make no mistake, GPT-3, ChatGPT, etc. will, for sure, mint a new generation of millionaires, but it very likely won’t be the random folks sitting on their couch eating ice cream while watching some random grifter’s content on YouTube. At best, the grifters will make lots of money selling courses (as usual) while the folks buying the courses will maybe break even. The real money will be made by the creative entrepreneurs and out-of-the-box thinkers who can come up with innovative uses of these technologies.
Create the out-group: "it very likely won’t be the random folks sitting on their couch eating ice cream"
Create an in-group: "The real money will be made by the creative entrepreneurs and out-of-the-box thinkers who can come up with innovative uses of these technologies."
And that's the key to fueling this machine. Someone out there just read that setup, said "That's me! I work hard and think out-of-the-box... I'll be the chosen one" then bought that hustle-bro's "Thinking Rich with AI" course they bookmarked from a Twitter ad last week.
Right, but that is true of everything. The real money will be made by the people who spot a business opportunity and exploit it. The desire to be one of those people will of course be ruthlessly exploited by unsavoury people, but nothing new there. If not for AI they'd be shilling about flipping houses, shorting stocks, etc.
I think many of the successes will be based around "inside the box" thinking now that AI is getting in the box. Speeding up tedious tasks, improving low-effort results, etc.
A non-tech one that was consistently pushed for a very long time and made a huge re-debut on tiktok (when it was popularized in the US) was dropshipping.
Just like you said, some people definitely got richer; Shopify, Amazon, and Alibaba to name some. Companies that built tools around dropshipping profited as well. The big names people throw around for why the idea is totally going to make them rich, those people got richer too.
Everyone else was slapped in the face with the reality that a lot of the big numbers they were fooled by were revenue, and just made someone who doesn't even dropship anymore $349 richer.
I remember watching an exposé of one guy in particular (I honestly can’t remember his name — it’s been years). He couldn’t prove that he had ever successfully dropshipped anything at all, yet he had raked in hundreds of thousands of $ via courses teaching folks how to dropship. Hilarious. I couldn’t tell if he was just a genius or if everyone paying for his courses were idiots . . .
I dunno dropshipping makes a lot of sense to me. Importing products from some cheap country and slapping your brand on it is what a lot of companies do quite successfully. I would rather categorize it with restaurants: Clearly and obviously a viable business model. But many fail.
Yup. I’m very familiar with the dropshipping shenanigans.
Let’s just say I have some family members who were 1,000% convinced they’d be multimillionaires from that and are now back working their working-class jobs, mortgage and all, no millions . . .
I've felt very frustrated by it too because it adds this very parsimoniously-constructed veneer of skill, and then makes it hard to filter out the signal after that.
I am a machine learning engineer. I'd like to consider myself a very good machine learning engineer for my age and experience, which is starting to get on in years as far as this version of the field is considered.
I am not an "AI Guru" or someone helping "Deliver Insights" when in reality they're just delivering slick marketing and a polished presentation of some sort. I do my research, I work on what I like, and what I'm good at, and what I need as well. If I'm able to do all three at once then that is good. Then I publish it, and if people like it, they can use it.
If it's good, the right people will find it after enough consistent word-of-mouth passing in the right places. I just find myself frustrated that people are just selling themselves and their identity instead of just selling their work or what they do from a very practical perspective, or putting it out there with appropriate advertising -- better yet. It feels like a kind of lying to me, and being on the autism spectrum, I'm more sensitive to that particular clash of values.
I also feel sad because not only does it devalue my career and the field that I like to quietly work in (though I do like celebrating my success when it happens), it also changes the expectations of my friends and family on the occasions that I do try to relate to them about the aspects of work that I like.
In a close vein to this, I saw an influencer this week whose primary line of work was optimizing gaming PCs for high-visibility figures in a viral way. That's it. That's all the influencer here was working on, practically. I'm sure there's a viable job path which includes setting up someone's gaming PC to perform optimally, but there is no way that this is a career. This is very silly to me, though I wish no major disrespect in my frustration.
I sympathize with you (the feeling of being threatened by charlatans). I don't think these feelings reflect reality, though. It can feel like your career will be threatened by the existence of this type of person, but unlike them, you have real technical skill to deploy on real problems.
> it also changes the expectations of my friends and family on the occasions that I do try to relate to them about the aspects of work that I like
While it can be grating that charlatans can have this kind of influence on the public's perception, realize that this is really the limit of their influence. They will never displace a person with real knowledge in settings where actually getting things done is the goal. Among your friends and family, it may even present an opportunity to practice your ability to translate your knowledge into something digestible to the layperson. You'll be doing them a service in the process.
Perhaps. Indeed it is true that my career is not threatened (as I would not want to work an organization that would fall for such a thing as this), but the thin plexiglass wall that it puts up is certainly real. People have their expectations, perhaps more than you might imagine here.
I put up the results for toy comedy prompt generation in a server that I quite like and got downvoted to oblivion (and it started a debate... unfortunately I think?) because it was using AI, because of course of some of the cultural associations with AI in some circles.
For me as well, it's given people an automatic "super smarty-pants person, I could never understand that" impression of me from the get-go by default or some other kind of impression that oftentimes drives a barrier in between people. In fact, my life has improved significantly by _not_ telling people what I do for work or by using a ton of indirection about what it is. A very significant social improvement -- people can more or less just take me as I am without that background perception running, thank goodness. It's pretty darn tangible in some circumstances too. It's much like the reverse racism that happens sometimes when one travels from a more "revered" country to another. Sure, one could "technically" get over that gap, but it's always there, and the emotions underneath still lie under the surface.
Another thing for example could be used to demonstrate that point is that I'm transgender (non-binary, to be specific). Before I knew I was trans (and even today, to a smaller degree), I'd feel extremely uncomfortable talking to a transgender person, even if on the outside I was very polite and kind. Of course, every person is a person regardless of their eyesight, hair, or gender, etc. And many people realize this in time. I've had a couple friends realize this a bit with me in my belief, and others vice versa to a degree as well I believe.
I hope that helps elucidate the problem. Where true technical skills are needed and the company is a good estimator of skill and usefulness -- sure, someone faking it in a sense is a good food-dye/food-colorant-like test of what my future would/will be like at the company with a string of incoming other coworkers. But there will almost always be places that sense and filter that technical skill well.
But where connection is concerned, they do and will have far-reaching consequences. Just look at StabilityAI, for example. A rapid growing startup with the mantle of appearing as an open-source, almost charitable non-profit initiative. But of course they're making the Faustian bargains for personal profit and growth at the expense of an industry -- and is interestingly enough why I got downvoted heavily in that one server earlier. It's also why I did not begin working with them in their infancy, though there a few other reasons of things not smelling alright within the organization that led me to that belief.
I hope that clarifies the issue somewhat further, I very very very much appreciate your comment and thank you for commenting very much.
In the video he talks about a course (which is a scam) selling a different scam of getting a ghostwriter to write a crappy book, getting someone else to record it, and then publishing it to audible to make money. which seemingly did work for a short period of time but no longer does. I guess this is the same sort of thing, except cutting out the ghostwriter.
I don't want to listen to unskilled voice-actors read a book to me. I definitely don't want some super awkward software voice that can't interpret the text to provide emotional color to their inflection. And this is not something I think is easily solvable by a 2023 AI startup.
that's why it's a scam. I think the original thing was that audible would give bonuses for publishing books (it's mentioned in the video, but it's been a couple weeks since I watched it), so you would basically just make shitty books to get bonuses. of course this no longer works, and now there's also much more competition, so your crappy grift books aren't actually going to get you any meaningful amount of money, but the point is just to sell people a grift they can do that seems like it could work
I wonder if we’ll move back to small static online communities (as were popular 2 or so decades ago). PhpBB boards with tens of regulars all of whom you recognize.
Agreed. I see internet culture shifting towards the fediverse, small chans, and boards.
Frankly it's too difficult to moderate at scale. And either way most users don't want to be part of the same big community as everyone else. They just want to be able to share content between their communities without a tonne of unnecessary friction or many different needless login credentials.
Twitter succeeded here because it allowed these communities to seamlessly blend into each other while allowing individual users to prune problematic individuals from their view of the community in a way that just wasn't feasible with other social media or forum platforms.
The fediverse (if the projects could fix their UX issues) look like they have the best shot of recreating this "small, discrete communities all intermixed" kind of environment. Hopefully however without a lot of the forced discovery and content rotating that Twitter currently pushes.
I guess next comes 'how will you maintain low scale'.
We've got to see the large scale problems in the last decade, but I think a lot of people either forgot, or were not alive for the small scale problems that existed back in the days of BBS's and early websites. The biggest issue with with most small communities is they tend to die quickly unless they have some unique bonding factor. For example a local RC car racing group. But that isn't generally interesting over a wide range of topics outside a few topics. So now you're looking at building multiple communities and managing them. But you'll find most people won't manage that well and they'll tend to gravitate to a few busier larger communities. But then those larger communities will grow much like the seed of a star via gravity, someone will know someone else that invites them and suddenly you'll have your own little facebook/twatter rife with all the same problems and split views on what is acceptable.
There are just a number of problems here that seem intractable when you have hundreds of millions/billions of people that want to be online and in communities of one kind or another, but an unending number of social problems that occur when you get large numbers of people gravitating towards interesting things.
I think small communities dying out is acceptable as long as they can be brought back by the members. That was one of the appeals of twitter for example.
You could get that tight nit group and maybe it kinda dies out for a year or two but something revitalized the community and it came back. With smaller, centralised communities this falls apart unfortunately since you have to rely on the host to maintain that service and if they shut it down, it's just gone. Federated services on the other hand have the advantage that as long as the participants continue to be part of that community, it can be revived on demand.
So in this sense, I think some of the major issues that prevented small communities from persisting and thriving have been solved. Of course like you mentioned the moderation, growth management, and UX aspects have yet to be solved but those have the advantage of being able to be largely independent of the core functionality and can be improved with time without breaking the world.
I also think the twitter approach to community formation works really well here. It has it's own issues but a lot of the more niche twitter communities tend to just fracture back into manageable micro-communities once they get bigger than a certain size. It's not as noticeable since what community you see depends on who you follow but it's 100% a thing that happens. I've personally seen it happen more times than I can count at this point.
Two solutions for intentionally keeping communities small and preventing low quality content are making joining "referral only" (like https://tildes.net/ for example) or requiring small fee entry (like https://forums.somethingawful.com/index.php) but each of those has their own problem. The prior has an issue of creating echo chambers as only "like minded individuals" can/will join while the latter can promote the exclusion of less well off individuals or those in regions with limited access to digital finance as well as posing a privacy concern unless some form of privacy preserving money is used.
Honestly I'm excited. Social Media from a technical perspective is finally getting interesting again.
HN is only this way because of a focus on moderation to keep the place of going off the rails. HN itself has one of the best moderators I've seen that's been able to keep out of petty arguments that tend to infect and ruin forums. It's not something easily scalable at all.
Part of the problem with being on a site like HN is we see everyone here bitching and groaning about how much 'big web' sucks and how we should get away with it, while missing that most of us here are also visiting that same big web with a few billion other people every day. I'll believe twitter/FB/etc is dying when I watch them turn off their servers for the last time.
HN is already way larger than that. There is nobody on HN who knows everybody else either by nickname or IRL or any combination. It's 'mostly anonymous' with 'mostly people I don't know or recognize at all'.
That's not exactly a very high bar, the problem is that it probably can also outperform your typical start-up founder and/or their corporate finance flunkies.
This is a fundamental shift that we're setting coming out of 2021-2022 and it's incredible.
I don't know what the next wave of the internet will look like but i'm actually excited because ChatGPT+DallE is making a whole space a commodity. People will need to learn how to adapt and unlike the previous shifts, it's not immediately obvious to me how things will net out.
This comment feels like a copy paste from a cryptocurrency/Web 3.0 comment from 2 years ago. It’s insane how cyclical the hype cycles are, and how they only seem to be happening faster and faster.
I mean if we're honest - it reads like a ChatGPT answer that shills itself. haha, but that's what's so crazy we're reaching the point where AI and humans aren't easily distinguishable and eventually impossible. Scary and yet exciting.
I'll be the first to pay tribute to our Skynet overlord.
I just don't see how chatGPT and crypto are at all comparable.
One immediately and obviously has thousands of use cases while one we are still trying to figure out the use case 13 years later.
Give it a few generations and I honestly don't know if I am reading this board anymore as opposed to just discussing things with the AI assistant.
If anything it is exactly why crypto is such bullshit. Something that is revolutionary doesn't need much convincing. If anything chatGPT reminds me of Myspace. All of the sudden everyone is just using it. No need for all kinds of childish and naive readings of Rothbard and Mises to convince people of its utility.
>and how they only seem to be happening faster and faster.
Eh, isn't that what the technological singularity people would say would happen?
Way back in the day I'd hear people go "oh the internet won't amount to anything, it's just a fad", and while I didn't know what the internet would become, I believed they were totally wrong. The internet was very useful.
Back in the very early days of bitcoin, back when I thought it was potentially useful but that people was highly confused if it was going to escape regulatory issues. When the bitcoin hype cycle started I knew people that knew nothing about blockchain were spouting dumb shit and that it would eventually end in a bust. Now, in my typical fashion I underestimated how dumb people can be on things like this and didn't keep the few bitcoins I had back in the early days and lost access to them missing out on a few hundred thousand dollars at the peak.
When looking at the different GPT-like models coming out I actually see useful products that can be made from technologies like this. For example there is one on the front page called 'universal summarizer' that can take a large body of text and turn it into a small and seemingly accurate summary. That is worth something for sure. The ChatGPT is presenting useful information to some of its users too. The biggest question about it's long term success is cost of operation. If that drops over time then GPT and whatever comes after that will be here to stay no matter what you think about hype cycles.
The bro addition is mostly sexist anti-male anti-community terminology designed to boost social signaling and communicate the orthodoxy's disapproval. Also holier-than-thou social psychology. Kind of serious, kind of joking here. Also a sign of the times where people who want to make money often get funneled into these types of very bleak pyramid games.
I've had similar thoughts, but this is the first time that I've ever heard anyone else talk about it. I wonder if there's an exact count of the number of different disparaging "bro" nomenclatures that we've had over the years.
I've already seen several videos talking about how you can use DALL-E or Midjourney to generate really samey-looking AI art and sell posters of it. This kills the whole art market by burying it under spam. Doesn't matter if you're drawing everything traditionally, using digital tools, or spending hours in front of Stable Diffusion to get it to spit out the image exactly the way you want. Your work will be buried under millions of hustlebros' spam images.
I recently watched a short video interview with an old music exec that claimed music production is dead. Now that virtually all music production is digital, and has been reduced to algorithms, new music will be created with near-zero time and effort. Bots will be able to crank out hundreds of songs per minute for any genre, then upload them to streaming platforms, flooding them and making it impossible for a human musician to compete.
He went on to talk about all it takes now is for public acceptance to catch up to the technology. He talked about how audiences used to demand real, live performances, not lip-syncing or playback tracks. Not anymore. Then people wanted actually singers, not Autotuned singing. Also, not anymore. He had a few more examples, but the point is that it's not a huge leap for the public to accept bot-generated songs; it's already accepted that nearly all popular music is digitally manipulated. It's only a matter of time before the public collectively says, "Why should I care if it's a computer or a human making the song? I just like it."
Will the AI generated music be lip singed by humans like Milli Vanilli? Because a big part of the appeal of music is the image and persona of the artist. So I don’t see pure AI music taking off until the entire artist can be generated. That means an AI generated human likeness with a personality that can be featured in AI generated music videos.
Either that or the metaverse takes off and avatars are normalized then I guess the artist could just be a computer generated avatar.
Everything will be AI-generated? Most vocals these days aren't really "human" anyway. Almost all vocals these days are digitally filtered (Autotuned), some singers/songs more than others.
So, yeah, that's the last step: detach the song from the singer, or the music from the musician. I don't like it, but I totally get it. The Gorillaz are pretty music there already. Yeah, they have human performers, but who are they? Where are they? I don't know. And I don't care. But I like it.
Or the performer is whomever you'd like it to be but some performers cost more than others based on rights? So if you'd like, you can be Rick-rolled by an AI generated dancing penguin in the style of Picasso with anime overtones for one price, but the original or AI up-scales thereof will cost you more (though, obviously, your willingness to accept AI-added background product placement built into the video will effect the cost).
The problem isn't the new technology being easier to "fake" the art. Concerns about some art being "fake" because it uses technology is just a weird form of artistic elitism: new, therefore bad. I, personally, have no problem with people using AI art (unless they're using it just to trace over other people's work, which is wrong).
My main concern is one of curation and discovery. The Internet already does a pretty bad job of these. And as a result it is very difficult for new artists to be discovered and find any sort of audience. But now you have a force multiplier that specifically allows people to produce plausible, trend-forward, but not boundary-pushing art with no effort. And you can churn out hundreds of these per day to completely corner public art spaces.
The downside of this is that the AI is statistics-driven, so it's going to reinforce a lot of the trends of "mainstream" art. Think about how a lot of AI art trends towards this kind of realism-plus, overly cinematic look. Anyone not making that is going to be crushed by the trends.
Which of course is already a huge problem in music. "Pop songs" are a very rigid formula, like to the point where new hit music often comes with a flurry of lawsuits because it's difficult to actually create legally distinct work in this genre. The reason why music works this way is that the people running these companies have the same mentality as the hustlegrind idiots telling you to go spam poster sites with AI art. They are not creating art to communicate, they are creating art to fill a market demand.
So my concern is the "popification" of art, writing, and code, as all of that gets delegated to machines that are constantly reproducing whatever was popular back in 2021.
"Then people wanted ... . Not anymore." I remember Elton John (singer known for energetic live performances) complaining at a music awards show in 2004 that Madonna won a best "live" performance" award even though she was lip syncing thoughout her show. No one cared.
My wife and I buy paintings. We either buy them from the artist or from galleries what we like. I don't really see how hustlebro spam images have any effect. I can see how CafePress type websites might be affected but at least for my taste I'm not worried!
I mean, it’s an eventual problem if an entire generation of aspiring artists disappear, because that’s where innovation and the cream of the crop come from.
Like, now, we have a hard time doing architectural restoration because going all-in on modernism destroyed the pipeline and ecosystem for artisans like stonemasons.
Its also a bad business to go in if you expect to make buck without being highly skilled.
Obvious artists are more extreme, but try getting a job as an aspiring programmer with no degree or experience who is just starting to learn to program. I imagine it would be about as bleak as being an aspiring artist.
Hustle bros were there since the beginning of time.
Think about that.
Every new thing for thousands of years someone always jumped on it for short term $$$ gain.
The hustle bros have a firm place in the ecosystem and the internet has just put a magnifying glass on the movement and made the hustle faster and easier.
Don’t hate the player (or the game), for they are as old as time.
"for they are as old as time" really isn't a justification for not disliking the player or the game. Shit is also around since the beginning of time, and we still generally avoid rolling in it.
>Nanni created the cuneiform letter for delivery to Ea-nasir. Inscribed on it is a complaint to Ea-nasir about a copper delivery of the incorrect grade, and issues with another delivery;[9] Nanni also complained that his servant (who handled the transaction) had been treated rudely. He stated that, at the time of writing, he had not accepted the copper, but had paid the money for it.
You can hate the player and the game. It’s not old as time. Capitalism and our current economic model as we know it isn’t as old as time.
Yes there were always hustlers and con artists, but they’re not exactly well revered. Jumping into a space and f*king it up for short term game doesn’t deserve respect?
Con artists before the modern era were reigned in by their local communities. A grifter would gain a bad local reputation and find themselves unable to continue running their scheme as they'd run out of rubes to dupe. They'd have to leave to the next town if they wanted to keep grifting. But now with social media, grifters can instantly find a new audience of rubes who are unaware of their past exploits
I think it was easier to be a grifter back in ancient times. Low communication among victims. No photographs to identify you , no permanant records (or at least not like the internet)
Capitalism is as old as time. People save resources, build wealth, and reinvest it. Happened in ancient Assyria, Greece, Rome, Persia, forever. Trade is an emergent phenomena.
The hyper restricted model of capitalism we have, where every last damn thing is regulated and taxed, is the innovation that began in Amsterdam and has continued for 400 years. But the concept of markets, trade, investment, and building wealth is as old as humanity.
>There are, it has been said, two types of people in the world. There are those who, when presented with a glass that is exactly half full, say: this glass is half full. And then there are those who say: this glass is half empty.
>The world belongs, however, to those who can look at the glass and say: What’s up with this glass? Excuse me? Excuse me? This is my glass? I don’t think so. My glass was full! And it was a bigger glass! Who’s been pinching my beer?
>And at the other end of the bar the world is full of the other type of person, who has a broken glass, or a glass that has been carelessly knocked over (usually by one of the people calling for a larger glass), or who had no glass at all, because he was at the back of the crowd and had failed to catch the barman’s eye.
This is where the most money is made; the structure is now and has always been before (at least since I have been old enough to notice it for the past 30+ years):
1. Make a product that gets some traction (make some noise, get some fake or real income)
2. Write, make videos, courses and products around and about how 1 became such a success and how someone can repeat this in 5 easy steps (which are pretty generic) and sell this for no, not $547, not $257, but a one time low of $167.
Number 2 vastly overtakes 1 in revenue immediately, but you need more 1 products to fuel the (sometimes snake oil) business of 2. So every new hype, you make a 1 product, use your list from the previous 2 product to make traction for 1 and repeat the cycle. Make 100k to a few million$ every cycle and keep an eye on ‘the next thing’ to make sure you are ready and early.
Many ‘open startups’ made (much) more money selling these 2 course/book products than what their startup actually tries to do. People want to make money doing nothing, not buy a subscription to your revolutionary ai powered todo app.
The funny thing is, about a decade ago in 2011, there was a trend in the art world called "zombie formalism", which was "a trend for market-friendly abstract painting that took the dead formalist aesthetics of mid-century abstract expressionists and brought them halfway back to life." It lasted for four years apparently.
Since then, an even wider movement that can be dubbed "zombie figuration" has taken up the art world in the last few years:
So it seems like many popular art scenes were somewhat engaging in creating their own deluge of derivative, samey art, even before this ensuing AI wave.
I still have some hope as I see some group of people started to tackle this space in different angle. Stop approximation as a product, wrap approximation around logic and output the exact result. ChatGPT or Avatar AI is literally approximation as a product.
The quality is objectively crap and the AI is not going to get much better. We've definitely hit the point of diminishing returns. What could possibly be next for generated content? It's not going to magically develop actual creativity emerging from mere statistical plagiarism. We're afraid of the literal average of everything?
Even as real human artists inevitably create cool new stuff, does it really matter that a bot can imitate the style later?
Also, if a "hustle bro" (dumbest word of the year?) can do it, why not everyone's dog too? This is just more pointless fearmongering. The covid years are over and this nerve doesn't respond anymore. They need to let it go already.
I'm sure around 1920 or so you could have said "Computers have never ____" and you could put pretty much anything in that blank.
Now the number of things that fit in that blank has shrunk significantly. If you do not believe the human mind is magic then there is very little that cannot go in that blank.
You don’t need to go back to the 1920s for that. I heard multiple experts in the 1980s say that if a computer can beat top Go players, surely they are intelligent like humans at that point. We keep moving the goalpost, but indeed that expectation for filling the blank goes up and yet it gets filled (and overshot by great stretches) a few years later.
:format(webp)/cdn.vox-cdn.com/uploads/chorus_asset/file/24403689/youtube_chatgpt_thumbnail_gore.jpg){kind=link}
Make no mistake, GPT-3, ChatGPT, etc. will, for sure, mint a new generation of millionaires, but it very likely won’t be the random folks sitting on their couch eating ice cream while watching some random grifter’s content on YouTube. At best, the grifters will make lots of money selling courses (as usual) while the folks buying the courses will maybe break even. The real money will be made by the creative entrepreneurs and out-of-the-box thinkers who can come up with innovative uses of these technologies.