5mo ago

Will AI Startups End Up Like Blockchain Startups?

In the ever-evolving landscape of technology and innovation, both blockchain and artificial intelligence (AI) have captured significant attention.

Let's think about blockchain for a bit. Blockchain technology has been met with considerable hype, promising revolution across various industries. However, this enthusiasm has not translated into success for most ventures in this space. Research indicates that approximately 95% of blockchain startups fail within a year of operation. Contributing factors include market volatility, regulatory hurdles, and the lack of clear use cases.

A notable example is the collapse of Terra's LUNA cryptocurrency in 2022. In just one week, $45 billion was lost, illustrating the inherent risks associated with blockchain projects.

AI startups are now experiencing their own wave of excitement and investment. However, they too encounter significant challenges. Over 80% of AI projects fail due to issues like insufficient market demand, operational difficulties, and ethical complexities.

Consider this: approximately 42% of AI startups fail because there is insufficient demand for their products or services. Not to mention, many AI ventures struggle with resource mismanagement, inadequate expertise, and scaling difficulties. You also have the additional challenge of navigating the evolving landscape of AI ethics and regulations adds layers of complexity that can impede progress. There's not exactly decades of history to refer to regarding legal precedent with AI.

A lot of the hype and marketing I see today looks just like what I saw a few years ago, except instead of "blockchain" it says "AI" now. There are consulting firms, integration firms, everything. Is this just a sign the industry is just endless fads with no actual commercial usage?

Bitcoin was hyped as reinventing the world's economy. Sure, it found a few usages, like replacing Western Union, or also by essentially becoming "digital gold" that people can just acquire and sit on, but last time I looked, VISA/Mastercard and the like were still doing 98% of the world's commerce. In other words, Bitcoin fell far short of where many of its proponents said it would land years ago. Looking around at all these AI firms, I wonder how many of them will even exist in 3 years.

56 comments

A lot will fail yes, but I think there's actually a lot of value in AI and many will succeed. Blockchain has always been a solution in search of a problem, but AI actually can help in a lot of ways.
- They’re putting micro chips in the cheese and using block chains to track the micro chips! https://www.businessinsider.com/edible-microchips-on-parmigiano-reggiano-used-to-fight-counterfeiters-2023-8
- A lot will fail yes, but I think there’s actually a lot of value in AI and many will succeed. Blockchain has always been a solution in search of a problem, but AI actually can help in a lot of ways.
  Well, what are those ways and when can we expect to see them? I keep hearing that "oh yeah, the very NEXT version of AI will do your job for you" but it always seems to be on the way. In the same way I can use blockchain tech for a few things here and there, I can use AI in the same way today. However, with all these billions and billions of dollars getting invested into AI right now, how will it change the average person's life in 1 - 3 years? I use that scale of time because that's pretty much how long startup runways last. If they don't turn a profit in that time frame, they go to the big AI graveyard in the sky.
  
  My wife uses AI to write complex Excel spreadsheet formulas saving hours. She still has to double check them but it saves enormous time. My friend uses it to write proposals. Again it needs to be checked and again it saves hours of time.
  AI doesn't replace people. It provides a productivity boost and it has been doing it for 2 years now.
  Asking what AI is going to do for the average person in 1-3 years is like asking what is the PC going to do for the average person in 1980. There's nothing that AI can do that can't be done by a person. But like using a PC instead of a pocket calculator, it makes you more productive.
  
  how will it change the average person’s life in 1 - 3 years?
  To be blunt, it'll probably make them worry about losing their job. AI can't do everything and it does a lot of stuff pretty shitty, but so do a lot of people.
  Tech workers, artists and other industries have long had to compete with work being sent overseas. Now it's even cheaper and faster.
  
  See, that's why this stuff is stumping people on all sides of the conversation.
  AI won't do your job for you in three years, or probably ever.
  AI is useful now, though.
  It's useful for specific stuff, when properly built into a process and mostly through more direct applications of machine learning than the firehose of generative AI corpos seem to think is a golden goose, but it's useful.
  There's a lot of chaff with the wheat here, a lot of people can't tell the difference and everybody wants to have an opinion more than they want to spend time understanding what's going on.
  So as always, a lot of people are going to lose a lot of money, a few people will make a ton of money and a bunch of narcissistic assholes will think that being the ones who won in that casino means they're infallible and should run the world.
Open ai made $2 billion on $5 billion in losses. Although, folks on here are talking about having it write proposals or spreadsheet formulas; this is not really a monetizable product. Aside from onsie twosie "it helps me write an email" scenarios, this isn't something anyone should be spending $20 a month to boil the planet for.
- Yep... and they are one of the market leaders. Imagine the margins on some of the other players and you get in the red pretty quick.
Yes, but not to the same extent. Both AI and Blockchain are amazing technologies, but those people that are pushing either as the next big thing since slice bread don't understand them.
Blockchain is an elegant solution to a decades old problem that's actually impossible to solve called Byzantine fault tolerance by making it costly to bad actors to the point where it's better for them to become good actors. It is revolutionary, but very unlikely that someone will make a profitable product out of it, especially because the two more obvious uses for it already exist and are open source.
LLMs, which is what people are calling AI, is also a very cool new step for text prediction. But it's not in fact intelligent, so it can't do anything without supervision, and more often than not it's easier and safer to create something yourself than to fix a possibly broken, possibly malicious creation by someone else. LLMs are great for stuff like brainstorming or suggesting short pieces of code that I was about to type anyways, but to think they can produce a book or a program on their own is absurd.
However, as much as I think Blockchains are elegant, they solve an abstract and very specific problem, whereas LLMs are good at solving generalized stuff. There are plenty of applications that would benefit enormously from having LLMs, e.g. a bot that finds, summarizes and points you to documentation at work would help anyone having to deal with documentation to make them more efficient, and companies that invest in these sorts of solutions might come up with great products. But most of the time they're using it as a buzzword or worse trying to remove a person's job which will backfire.
- bitcoin is a workaround, not a solution to the byzantine problem. Its security hinges on humans placing value on it and wanting to hoard it. It's a clever psychological thing that makes it so that the participating cleanly is more advantageous than trying to cheat. But the byzantine fault problem is just as mathematically impossible to solve as ever (because proven math does not change).
  
  Absolutely, I even mentioned that it's an impossible problem, even so it's an elegant solution to it, because it's more profitable to participate.
- Blockchains solve nothing, lol
  Now blockchain fans need to come up with problems it solves, fucking Byzantine fault tolerance, lmao
  
  Is it that you don't understand the problem or the solution? Or maybe you got scammed and lost money by buying a picture of a monkey that someone assured you would be worth way more, akin to an old person being against emails and thinking they solve nothing because once they sent money to a Nigerian prince.
I think you rediscovered (1). Yes, most of those startups will disappear, pivot to the next hype, ...
- I was racking my brains for hype that goes further back... before blockchain there was mobile phones / apps getting hyped (although the whole world DOES use a phone, so I guess there's that) and then before that was web, but I don't think either of those bubbles were quite as insane as blockchain / AI in terms of "what they promised vs. what we got".
  
  Self driving cars, social media, 3D printing.
  It does seem to get louder and more agressive each cycle, yes. But then again, all media is.
  
  Old man here... The first online bubble was probably the dot.com bubble of the late 90s, when lots of people first went "This internet thing is amazing and we're all going to make millions!" I remember boo.com being one of the first high-profile crashes, but pretty soon a lot of that first wave of internet retail businesses folded, with notable exceptions like Amazon, of course.
It's less of a bubble than the block chain - by a really wide margin, the tech itself is already more useful itself than the block chain ever even pretended to be moving towards - but it IS still a bubble.
Some are grifters and will move to the next fad, some simply don't have good enough ideas and will fail, others won't be able to achieve their good ideas as funding slows.
Well yes and no. Startups will die because many do anyways but AI technology is not failing. As an IT professional I can tell you that companies are investing in AI power more and more and, as a regular employee, I can not explain how much time I save using chatGpt alone. BUT... as in google translate, if you know nothing about the subject and is relying only on chatgpt you may have a bad time.
- How dare you say something positive about LLMs on Lemmy? At least 2 downvotes, people here are a joke.
- One significant red flag is that GPUs that were selling like hot cakes because crypto are now selling like hot cakes because AI. It leads to a proliferation of gold diggers buying pickaxes instead of a refinement and optimization of the underlying technology.
  Paraphrasing Aldous Huxley, “the next tech bubble must need at least as much GPUs as the last”.
Fingers crossed.
Yes
Most AI startups are just middle men to OpenAI. I don’t think these startups will endure. People rather just use OpenAI directly.
I think there’s a future in AI. It’s good enough for many people today, and it will continue to evolve. Like it or not, we’re going to see more of it.
Blockchain is just MLM snake oil. The reason it’s still around is that it’s a great tool to separate people from their money.
- The reason it’s still around is that it’s a great tool to separate people from their money.
  It's true -- it's basically made a Ponzi scheme 50x easier to run (and with global reach). Not the kind of advancement I would have wanted, but it's the one we got...
too much grift happening in both... whatever economic value both technologies were gonna provide has been too muddied now.
I think you're so close here, you gotta look further at the overlap. The venn diagram of crypto bros and AI bros is a circle. What's happening is a constant cycle of out-of-touch business execs and fresh MBAs thinking they can print money using the hot new scientific breakthrough. Same as the Dot Com bubble, it's all just new snake oil for the salesmen.
- I did notice how many "crypto influencers" are conveniently re-branded and not selling NFT anymore... they are all selling things like "Improve your business with AI! Take my course!"
tldr
they will both fail, yes
Hopefully, yes.

56 comments