The real engine behind "AI" is the bad tech economy | Hacker News Reader

The real engine behind "AI" is the bad tech economy

hachyderm.io

80 pointssoftwaredoug73 comments

Not really sure what exactly the author is getting at but let's be real: the IT industry is awash with investor dollars hoping to fund the next facebook. AI enters and it starts doing actual, useful work, and now there are a zillion VCs dumping out their wallets trying to get a piece of the next facebook.

None of this means that AI is dumb, rather, VCs are dumb and will chase literally anything that smells like money.

Are they dumb or are they good at investing in stuff on very favorable terms for them and then pumping and dumping their portfolio of hype companies on the unsuspecting masses at a major premium to their initial investment?

Can be both dumb and "good" at investing.

>If you're so rich, why aren't you so smart?

This is a question I wish people would ask more often, but instead, determinism has people associating wealth with merit, and even worse, absence of wealth with absence of merit.

Independent of the "real" value of AI

My argument would be the sudden pivot of EVERYONE to AI is more indicative of a lack of good prospects for investments, rather than companies actually discovering products users want and like.

Of course there's real value, but I'd believe the hype more in a strong tech economy, and am distrustful of every company becoming an AI expert overnight in our current weak tech economy.

But this is how it's always been. Someone could have funded AI heavily (and a few did) during the web 3.0 hype but most went with the crowd. Same with streaming video services, social media companies, hardware companies, MMOs, search engines, cloud, etc.

... and just prior to the big attention-boom for AI many of the same players were trying to make their variant of the "metaverse" the Next Big Thing -- most of them advocating for the importance of virtual meeting spaces while simultaneously mandating RTO.

> indicative of a lack of good prospects for investments

This is what it looks like to me, too. Also, all the money that poured into "crypto" enterprises and schemes. People are inventing $THING_TO_INVEST_IN out of nothing, and capital piles in.

I'm no economist, but this seems like the result of there being too much investor money chasing too few investment opportunities. Why's there too much money? I'd speculate that QE caused this.

I've always thought that QE was nuts; the present global inflation scenario was entirely predictable.

And i wouldn't even say VCs are dumb because they keep hitting success chasing that money. It's why they have this money to throw around.

The labor market is USD$144 trillion worldwide in salary. Automate a fraction of a percentage of the worldwide labor? You have a company with more annual revenue than Apple.

Is it that hard to understand? Can people not think of scenarios where the current generation of AI may be able to automate some percentage of jobs? eg. Do people have any idea how much is spent on paralegals looking up case study law and how good LLMs are at finding such links in an automated way?

I keep seeing people with no idea of the market pushing bad ideas. "LOL CEO said 'AI' to make stock go up". "Nvidia can't possibly be worth more than it once was because AI isn't real". "It's hype and doesn't work". etc.

only if they can keep training from the commons and stealing other people's work.

Even looking at a world where training data, models, etc. are all public domain or otherwise free to use for commercial purposes, you will still see a huge number of companies that specialize in AI "as-a-service" simply due to the size of the ML talent pool, the desire to mitigate risk and purchase a known quantity rather than "roll your own AI", and the nature of how new technology is implemented at scale.

Edit: and if copyright laws strike down current AI models and require some kind of BOM with established 'rights' to the data used to train models, you'll just end up with some variant of a gray market for training data that allows big corporations to do what they want with whatever data they can get their hands on, delegating risk to middle-man brokers making access to trained models EVEN MORE restricted to highly funded ventures.

Pretty soon, we'll be forced to wear glasses so that companies get revenue when we look at something copyrighted.

I understand copyright holder's complaints. Maybe we can meet in the middle - reduce copyright to 5-10 years.

Indeed, the current trajectory of discrediting the public domain leads towards a pay-per-thought society.

What is being stolen, exactly?

Knowledge

> The labor market is USD$144 trillion worldwide in salary. Automate a fraction of a percentage of the worldwide labor? You have a company with more annual revenue than Apple.

What makes you think that AI/LLM tech is going to result in a single winner? That revenue might be split between a few hundred thousand companies.

Unlike the product from the current tech giants, there is no network effect AI stacks, there is no lockin, and there is absolutely no way to build a moat that can't be crossed with minimal capital.

This probably why Google sat on this for so long and refused to make any progress forward. It took openai to take Google's research and actually implement it. Google weren't going to do it because there's no moat to speak of

I don't think that's why Google sat on it. My guess is that it was hubris of the key researchers. They all had these castles in the sky of how it was "supposed to work" and then all that is required for quite powerful systems is scaling.

Only Ilya seemed to lack hubris and followed the path of scaling.

As for their being no moat, we don't know that. We are seeing a number of attempts at replicating GPT3/4 and there has been some success which is indicative that there is no moat, but this is a really new field.

> That revenue might be split between a few hundred thousand companies.

That’s not how power laws work. Over time, many niches will be created but usually it starts with one big winner taking it all. For example, first there where many half baked solutions, then there was Netflix and now we have all kinds of streaming providers, many of which are trying to find a niche such as Apple-owners or Disney fans. Similarly, we first had an iPhone and now small phones, big phones, tablets, touch laptops etc etc.

> Over time, many niches will be created but usually it starts with one big winner taking it all.

Usually, yes. But in this case it appeared to have started with multiple small providers, with multiple non-half-baked services. There doesn't appear to be a moat for anyone in this market.

Unlike streaming video (Netflix, et al) which has a moat (exclusive content), or social media (network effects), or ad industry (eyeballs).

With AI there is no moat that I can think of (quite happy to be provided with potential counter-examples, even if they are only hypothetical).

Let me try a different way to explain it: If you're building an ad network, a social media service, a SaaS dev-product provider, a video streaming service, etc, you will be unable to convince users to switch just by having a better product. The moat will keep them on the worse product (eyeballs, network effect, legacy support, exclusive content respecitively) no matter what you do.

With AI, there is no moat that I can think of - if you were to start an AI product today, what is the AI-specific moat you can use to keep out future competitors? Legislation, maybe?

I'm not sure that iPhones are a good example. It's a product, not a service. By definition there's no moat.

> With AI there is no moat that I can think of

I was also thinking about that yes. I think it’s mostly data (a sort of network effect), capex, and state-of-the-art engineering/research. A bit like TSMC and ASML. Also heavy capex and engineering moats.

There may be many dumb VCs, but both Mark Zuckerberg and Paul Graham have more money than I ever will. Whatever they're doing - and in Mark's case it might be evil - I wouldn't call it dumb.

People have short memories. A few years ago hypes like big data and machine learning, blockchain and autonomous cars were pushed relentlessly. The consultants would confidently predict 300 gazillion profits by 2030, everybody would lose their job etc.

Then the pandemic happened, momentarily some candid voices admitted that all this digi-crap does not help an iota with our quality of life (or life, period) but the digital WFH hype found just the right meme to sell.

Then the pandemic receded but the LLM breakthrough was compiled just-in-time to rekindle the AGI hyperventilation...

There is no respite. The tech economy is like the fabled sharks: needs to keep inventing hypes or it will implode.

If I had a dime for every conference I went to from 20010-2015 that said "IoT will be a 20 trillion dollar market in 10 years", I'd be retired.

I can't remember the name of the company, but they sponsor CPR(Colorado-area NPR affiliate), but they went from being "an IoT company", to "an IoT blockchain company", to an "IoT blockchain with AI" company.

C3 IOT I'm guessing

This is the one. C3 IOT became C3 AI without batting a lash. The radio copy is almost identical, with only the name change to catch your ear or avoid sounding dated.

That is really funny. I wonder what it is like to be that audacious?

I think you got your timeline a bit mixed up. the EV and blockchain craze happened during and BECAUSE of the pandemic. It was the rise of inflation that brought everything down starting in February of 2021. (yes, blockchain had a previous bubble, but that ended in 2018).

This is why bitcoin and the nasdaq both peaked in December 2021. But those EV spacs were already in a world of hurt since Feb.

Nah, the hype cycle for both started a decade ago, long before the pandemic.

Big data and AV were overhyped buzzwords at the time, but they’re both actual technology that are slowly reaching normalcy.

Plenty of people are using Spark and Databricks, plenty of people are using more abstracted products like Bigquery. No, people developed the good sense not to try to run clusters of janky and poorly documented Apache big data software as if it were “free” (except in maintenance… and bugs… and lack of feature support…) but big data is absolutely a thing. It’s just lost its buzzwordiness in the same way Web 2.0 has.

I’m actually surprised that the AV hype has died down so much because there are actual L5 AV products (Waymo, Cruise) regular people (in SF) can use now. We’re in like the original iPhone stage. They’re having their moment.

The biggest lesson I took away from all of these hype cycles is how unpredictable they are in their arc. Some trends like AV seem like they were going to take over the world, but they died down as quickly as they rose when we realize how hard safety is.

VR seem like it was dead with Google Glasses and Meta. But recent releases show that the [vision of VR is more real than I thought](https://www.youtube.com/watch?v=MVYrJJNdrEg). Apple releasing their VR definitely provides more legitimacy to the idea.

Technology is always hard to actually implement in the real world. I think have expectations that the world would flip over with new innovations in a matter of months. But it actually takes years of hard work to get people to adopt it. The internet and mobile phones are good examples where the promise was clear but the journey being universally adopted was not easy. *

* For the internet, you have the massive struggle of web standards, getting implementations web browser to align, and achieving acceptable performance. Internet Explorer still haunts the dreams of many web devs. It's remarkable how many fully featured web apps there are like Onshape and Google Maps.

* For mobile phones, they require a massive amount of technology investment and unit economics to get the price cheap enough for normal people. It's amazing how wireless technology gotten to the point where we can stream videos on phones.

I think the time to value between the Internet’s hype cycle and real utility was relatively short compared to most other technology hype cycles, because it was like the OG computer hype cycle and so was naturally a bit more delayed than subsequent hype cycles that had as their rationale “better not miss out in case this as big as the Internet”. It only took about 5-10 years to go from the beginning of the bubble to the modern incarnation of the internet and early versions of the products that now underpin trillion dollar businesses.

But, a lot of internet standards had been in the works for a long time at that point. Not so for AV and VR. I think we’re looking at 2+ decades for a lot of technology like that. It’s such a long time it really shakes out the hypemen with short attention spans.

Yes because the past data is so good to predict the future.

I guess technical analysis isn't that bad!

In shot form -- The same people that heavily marketed Crypto and NFTs (Which didn't work as promised) are now the same people over-marketing deeply flawed Ai tools for monthly subscriptions.

You can literally develop a front end for an Ai tool and connect it to a social media scraping back-end service in a day, and then hide it behind a subscription pay wall. This is what is driving the flood of tools right now, not well thought out and unique truly Artificial Intelligence-driven projects.

> momentarily some candid voices admitted that all this digi-crap does not help an iota with our quality of life User behavior was completely inverse, consumption for everything tech went through the roof. H/w for WFH and home schooling, Netflix++ for entertainment, everything ecommerce.

Everything follows hype cycle. It is just that tech is moving so fast that you see new cycles every 5 years. But apart from crypto, everything has delivered.

> The tech economy is like the fabled sharks: needs to keep inventing hypes or it will implode

No, the tech economy wouldn't implode. It would just signal it had become an old, stable industry (eg automobiles). The so called tech economy isn't going away under any scenario, it underpins modern civilization across the globe, and that's only going to be more true in 20/30/50 years than it is already today.

The tech economy is a thousand times larger than it was in the mid 1980s. Despite one hype cycle after another. If your premise were correct, it would have ended a long time ago because it would all just be hype and it would inevitably collapse in on itself.

You've confused a necessary ingredient: venture capital investment, speculation, experimentation, some chaos, with thinking nothing good comes out of the process. The exact opposite is true, it's that very process that produces drastically more positive income and wealth than is destroyed in the inevitable pull-back of the hype cycle (whether it's gentle or a dotcom bubble style crash).

Should take a person less than a minute to grasp this, if they have any awareness of the extraordinary size of Apple, Microsoft, Google, Facebook, Nvidia, Taiwan Semiconductor, Amazon, etc. versus the dotcom bubble days (for the equivalent style companies of the time), much less the 1980s.

The global economy (as well as the US economy) is now drastically larger. If you produce a successful product, your outcome can quite easily be drastically larger accordingly. The global economy has expanded in a truly massive way over the last several decades, and the tech economy has gone with it. It's not riding on false hypes, generative AI and the like are a itsy bitsy tiny little corner of the present scale of the tech economy. Go take a look at Microsoft's operating income over the last 30 years and tell me what you see.

> Should take a person less than a minute to grasp this

Well, we need to calibrate a baseline otherwise its just a waste of my precious minute. Digital tech as such, especially the hardware side is not a fluke. It evolves inexorably in capability, follows physical laws (e.g. the endless shrinking of transistor sizes has stopped) and is the tooling supporting a range of potentially long-term sustainable economic activities.

This reality does not exactly fix the relative size of the tech economy to the rest of the economy, much less so in terms of market valuation (which is the driver for most the actors involved).

There are various important factors that people simply assume are "normal" in the tech landscape but they are not. They are the result of arbitrary political decisions. 1) The degree of allowed concentration (monopolization) that essentially guarantees rent extraction and prohibits commoditization. 2) Lax attitudes towards data privacy that effectively simply siphon wealth from others. These two regulatory malfuctions could be reversed without much impact on the actual tech but a major impact on valuations.

In fact the engineered hypes serve as much the need to keep inquiring minds at bay (by arguing now is not the time to stifle the exponential innovation) as they do to inflate stock options.

Finally you have the fact that a certain fraction of "tech" is essentially disemboweling and substituting (rather than growing) other industries. For each triumphant Amazon type entity (that you seem to celebrate so much), you have the millions of boarded-up city bookstores and shops. This is a more complicated economic calculus. You could argue that all those vendors, publishers, taxi drivers, hoteliers etc should have been more proactive in adopting technology rather than have their lunch eaten.

I would like to reverse the compliment and argue that it should take a person of any awareness less than a minute to grasp the rotten foundations on which quite a bit of the "tech economy" has built its edifice.

> big data and ML

Became a big thing after Amazon predicting someone is pregnant before they know it, hype died off after actual implementations predicted that people drink coffee in the morning.

The hype died but those with actual use cases are using it.

> blockchain

Became a big thing after it made crypto millionaires, died off after people figured out that it doesn't help with anything other than gambling and illegal operations because its not better at anything that has an established legal infrastructure

The hype died but people still need the gambling and hope to be rich on the next hype cycle by jumping on something like NFTs.

> autonomous cars

Tesla, although never delivered the thing they promise, it was able to propel them to the top as a car company. Now Tesla is a serious brand.

Mercedes actually delivered autonomous driving car you can purchase and others actually were able to offer self driving taxi services but nobody seems to care that much. For some reason, Musk won this one without actually delivering it.

The real engine behind “internet” is the bad tech economy.

Vendors need something to sell, companies need to show results to investors, team leads need to justify staffing, devs need to stand out to avoid layoffs.

Insert any buzzy topic into this mix it will get eviscerated by the starving packs of tech jackals… regardless of what actual users want.

I’m excited about the potential behind the technology. But we’re in a state where many emporers have no clothes, and nobody quite wants to say it.

Well "Internet" went through its own boom and bust crash during the dotcom era. With lots of dumb money chasing dumb investments.

My hypothesis would be that cycle is going to be very heavily accelerated due to the current economy. Many companies aren't operating from a position of strength, but more of desperation. Some tiny few will do well. Many announcing their "AI thing" suddenly becoming AI companies overnight will probably crash and burn.

I could be wrong though.

Your thoughts are valid. AI.com is the Pets.com of this generation

We are in a 2nd bigger 'tech' bubble. I hear lots of funny thing said about the era and now' "today is nothing like the dot com bubble, pets.com ... PETS DOT COM!". and I am like, pets.com had a max valuation of 400 million for like a hot minute. Today we have over 1,000 never made a profit tech unicorns. One thousand companies with a market cap of 1 billion+ that has never (and probably never will) made a profit.

"you just got to find a company with a strong balance sheet, one that makes money, one that sells the shovels". Had you invested your money in CISCO back then, you'd still be in the red over 23 years later.

Geees well when you put it like that you make it sound pretty bad

The biggest losers on the dotcom crash were actually the failed networking companies "selling shovels" who had many 10s billions invested and then went to zero. Only AOL/TW was really comparable to Worldcom, Globalcrossing, with the $100B failures of Nortel and JDS/Uniphase.. Pets, Boo, Flooz etc made a big ad splashes, but were actually pretty small (single digit $B).

There have always been naked emperors, but what has changed is the amount of pretend time + energy in the form of currency that society has been willing to produce, and pretendess is the one thing that feeds more pretend ess. The pipeline of pretendness is ignorance, both willing and institutionalized. In prior times, the rubber band would have snapped by now. The fantasy has become so hyper that it seems nobody has a clue whether there is even a "rubber band" at this point.

There was a desperate attempt to find something, anything, which was a "must have" item. This fueled the VR boom of 2019 and the metaverse boom of 2021. Both sort of worked technically but turned out to not be that compelling. It's not yet clear how the "AI" boom works out.

It may just be too soon. VR is stuck because the headgear remains clunky. The metaverse is stuck because the initial implementations were crap and you need a serious gamer PC to really make it work. AI is now running into the problem that large language models are expensive enough to run that they can't be ad-supported. All three may pay off as the technology progresses.

> VR is stuck because the headgear remains clunky

It could be stuck for more than one reason. I would find VR repugnant even if it were as easy to wear as a contact lens.

That may be a generational thing. There are "VR natives". See Phia's Virtual Reality Show.[1] There was a time when people didn't walk around with little screens in their hand. There was a time when only women used keyboards.

[1] https://www.youtube.com/channel/UCMMBpWfWUd3xlcOxrot_neA

Agreed. My own personal trajectory was

1. raised without Internet

2. discovered internet and thought it would transform the world for good

3. discovered mobile devices and thought they would transform the world for good

4. discovered social medial and thought it would transform the world for good

5. deeply disillusioned with social media and abandon it

6. think social media festering cancer

7. think mobile device cuts me off from most meaningful social contact

8. abandoned mobile device (but kept Apple Watch for emergencies)

I'm curious to know if future generations who started at #4 end up at the same place I did.

> AI is now running into the problem that large language models are expensive enough to run that they can't be ad-supported.

You mean the 3 large companies that tricked everyone into handing over their data and processing capability to the "cloud" have almost reached a point of diminishing returns due to the amount of processing they are trying to do on it?

Oh the humanity.

Oh well. Maybe stop post-processing everything?

I think it's time to design an open-source federated processing platform where we amortize costly processing tasks to large batches of devices. We have never had more powerful computers, and less control over code or data.

I pitched something like this to YC years ago. I guess I should have known better than trying to pitch a cloud computing alternative to a bunch of VC's who regularly tell people that infra is a liability. Maybe if I started my pitch with "rent a bunch of AWS services and..."

This is a dumb take. While there is plenty of BS and vaporware masquerading as "AI", there are plenty of real companies and startups working on leveraging "AI" within their own domains.

A business isn't just technology, it's also about the business processes developed, and plenty of AI startups are honing out processes that larger companies will either scoop up, or that these companies will be able to grow up and compete.

AI is a marketing term. The people using it are marketing something. You would never use the term in a technical description because it's too ambiguous. Everything he's describing: Companies showing off to investors, devs showing off to management is about marketing.

Yeah for real. Saw a post on HN yesterday I think with the title something like "Development we call it machine learning, in advertisements it's called AI" or something to that end

But that is a good! Contrast that with other countries where the financial landscape is super conservative when it comes to investment and innovation, like in many EU countries. Economic growth is lower, jobs are fewer, and salaries are lower, and there are less exciting jobs where you can truly build the future. Americans do not appreciate what they have. You do not want the alternative! Appreciate your overly optimistic dumb money. It is a good thing for everyone.

And yet, despite having less jobs and lower salaries, many countries of europe tend to have less signs of poverty: women are less likely to die in childbirth, there's less gun violence, less homeless and child hunger...

The US has half the homelessness rate of the UK or France. Childbirth is due to lack of universal health care and gun violence is due to the massive number of guns – not the tech economy.

That might be right, but bear in mind the UK counts temporarily housed people (e.g. people couch surfing) as homeless, not just people literally on the streets. These stats may not be directly comparable.

As someone who wanders around the US and UK a bit, they're definitely not comparable.

Those are orthogonal, or are you implying lack of VCs create those conditions?

> women are less likely to die in childbirth

This one has been debunked many times. US healthcare attempts riskier births that would otherwise be aborted.

What I see is that the underclass in Europe is better taken care off than the underclass in the US. But the working and middle classes in America are richer and have more opportunities than their counterparts in Europe.

All of this, is completely orthogonal to whether or not private capital is willing to invest in risky ventures rather than hoard their capital in very secure incremental investments.

I don't want to build the future I just want dental care for my children.

Universal health care is important, necessary, cheaper than our current system, and I support it. But it's a completely orthogonal concern to the state of the tech economy.

I see an argument in this thread for laissez faire capitalism; its political ramifications are entirely relevant to the state of universal health care in the United States. The economy is interconnected -- the mathematical concept of orthogonality is rarely appropriate when describing broad economic complexities. But perhaps you should simply suggest that a health care discussion here is off-topic?

Culture might be a link between risk taking venture capitalists and less publicly financed social security. However I do not think there is an economic lock that forces a society to be either or. I can easily imagine a society that provides both a good safety net and where capitalists are willing to invest and innovate.

I guess what is missing is the right mentality and attitude, so like, our cultures have not evolved to that level yet.

I grew up in (western, to be clear) Europe and never received adequate dental care either.

I'm as much of a skeptic on "AI" as any. It's important to stay grounded and realistic to not unnecessarily create fear.

But you can't toss aside the entire space either. It's clear there are many useful ideas and improvements. For example, Retrieval Augmented Generation (RAG) when done right is ready for production apps. Vector search clearly can find matches that BM25 can't. Generative text-to-image models are enabling many to be creative in ways not possible before.

The last year+ has produced tools when used right can be quite useful to many.

What does he mean by "bad"? What would a "good" tech economy look like? Hard to define. Easy to misinterpret and argue over our different understandings of what he meant.

"bad" means investors want to make money on their investments, and tech workers want to advance their careers, I guess.

I would define a "bad" company as one that puts profits over people. People includes customers (incl b2b) and employees. A "good" company is one that realizes treating people with respect is a long-term profit driver.

The way you've phrased that, the good companies still put profit first, it's just that they have a different strategy - treating people well to build loyalty etc.

> What would a "good" tech economy look like?

Producing goods and services that real people buy and benefit from.

Low effort post. An AI probably wrote it because it’s full of meaningless platitudes.

I think the fact that OpenAI fired the starting gun is a bigger factor than an industry wide conspiracy.

Reply on news.ycombinator.com