AI Startup Boom Raises Questions of Exaggerated Tech Savvy (wsj.com) 51
SoftBank-backed startup offers 'human-assisted' artificial-intelligence; current, former employees say company inflates its tech expertise. WSJ reports: Startup Engineer.ai says it uses artificial-intelligence technology to largely automate the development of mobile apps, but several current and former employees say the company exaggerates its AI capabilities to attract customers and investors. The competing claims reflect a growing challenge in the tech world of assessing a company's proficiency in artificial intelligence, which refers to technologies that can allow computers to learn or perform tasks typically requiring human decision makers -- in many cases helping companies save money or better target consumers. Because AI technology is complex and loosely defined, nonexperts can find it hard to discern when it is being deployed. Still, money is flowing into the sector, and many startups can say they use AI as a way to lure investments or corporate clients even when such claims are difficult to vet.
London and Los Angeles-based Engineer.ai raised $29.5 million last year from investors including Deepcore, a wholly owned subsidiary of SoftBank. Other backers include Zurich-based venture-capital firm Lakestar -- an early investor in Facebook and Airbnb -- and Singapore-based Jungle Ventures. Engineer.ai was spun out of an earlier company in 2016, the company has said. When announcing its funding last year, it said it had notched $24 million in revenue while self-funding its operations. Engineer.ai says its "human-assisted AI" allows anyone to create a mobile app by clicking through a menu on its website. Users can then choose existing apps similar to their idea, such as Uber's or Facebook's. Then Engineer.ai creates the app largely automatically, it says, making the process cheaper and quicker than conventional app development.
[...] Documents reviewed by The Wall Street Journal and several people familiar with the company's operations, including current and former staff, suggest Engineer.ai doesn't use AI to assemble code for apps as it claims. They indicated that the company relies on human engineers in India and elsewhere to do most of that work, and that its AI claims are inflated even in light of the fake-it-'til-you-make-it mentality common among tech startups.
London and Los Angeles-based Engineer.ai raised $29.5 million last year from investors including Deepcore, a wholly owned subsidiary of SoftBank. Other backers include Zurich-based venture-capital firm Lakestar -- an early investor in Facebook and Airbnb -- and Singapore-based Jungle Ventures. Engineer.ai was spun out of an earlier company in 2016, the company has said. When announcing its funding last year, it said it had notched $24 million in revenue while self-funding its operations. Engineer.ai says its "human-assisted AI" allows anyone to create a mobile app by clicking through a menu on its website. Users can then choose existing apps similar to their idea, such as Uber's or Facebook's. Then Engineer.ai creates the app largely automatically, it says, making the process cheaper and quicker than conventional app development.
[...] Documents reviewed by The Wall Street Journal and several people familiar with the company's operations, including current and former staff, suggest Engineer.ai doesn't use AI to assemble code for apps as it claims. They indicated that the company relies on human engineers in India and elsewhere to do most of that work, and that its AI claims are inflated even in light of the fake-it-'til-you-make-it mentality common among tech startups.
You don't say (Score:4, Funny)
Re: You don't say (Score:2)
It's too bad all the hydrogen for fuel cells got used up in fusion reactors instead. We could have had fuel cells in our 3D cars by now.
Re: (Score:2)
But what if we used "blockchain" to help our "ai" help our "developers" create your "app" to earn you "money"?
Those buzzwords?
Re: (Score:2)
Re: You don't say (Score:2)
Hopefully it's not the buzzwords, but the hundreds of incredibly useful apps that will inevitably come.
I remember when I made my first AI-assisted program 20 years ago. That syntax highlighting really sped the process along. I could do the work of 1.05 men!
By the way, I am willing to relocate, as long as I can be guaranteed a certain amount of shares in the IPO.
Not surprising (Score:4, Funny)
You're not going to get funded unless you're selling AI-powered, ML-enhanced, blockchain-enabled, web-scale serverless functions in multi-cloud platforms.
Is it a shocker that you have people piling into the market trying to paint anything with the AIMLBlockchain brush?
Re: (Score:2)
You're not going to get funded unless you're selling .....blockchain-enabled,......
Is it a shocker that you have people piling into the market trying to paint anything with the AIMLBlockchain brush?
Seems to have not worked out for Long Blockchain (formerly Long Island Tea Company). They got themselves delisted from NASDAQ and an SEC investigation.
Re: (Score:3)
It worked out fine. The stock rose 400% after they made the announcement and the founders sold out. The company was going to go out of business anyway, so being delisted was fine. It was just an exit strategy. Maybe the SEC will give them a small fine, but they will likely come out ahead.
Re: (Score:3)
You're not going to get funded unless you're selling AI-powered, ML-enhanced, blockchain-enabled, web-scale serverless functions in multi-cloud platforms.
. . . if you throw in IoT, the Ginsu and the Spiral Slicer, I will invest in your startup.
. . . but wait! There's still more!
Re: (Score:2)
An AI is an AI is an AI (Score:3)
"They indicated that the company relies on human engineers in India"
So AI means Asian Intelligence in most cases.
Please read and ... (Score:3)
... apply to quantum computing.
I preach and they don't listen.
Re: (Score:2)
The best thing to do is to short all of these hype companies. Unfortunately Dwave isn't a public company.
Description is inflated, too (Score:1)
"which refers to technologies that can allow computers to learn or perform tasks typically requiring human decision makers"
I've come to the conclusion that AI isn't this at all, but rather it is just pattern matching. AI could generate code that matches the pattern of other mobile applications, but it really doesn't "know" what it's doing, and can't make knowledgeable decisions about how that code would differ from the code it "learned" from, in order to accomplish whatever the new mobile application is sup
Re: (Score:3)
This is why I vastly prefer the term "Machine Learning" instead of "Artificial Intelligence", because I think it captures the general essence of what most neural-net type algorithms are doing far better. As you said, it's just advanced pattern matching, with humans carefully selecting the criteria for evaluation and training the nets themselves.
That doesn't mean there aren't impressive advancements in the field. For instance, cars are driving themselves quite effectively now and should be out on the mass
Re: (Score:2)
That's not entirely true. There are lots of techniques that don't use explicit cost functions the way simple supervised learning does. Most of what you hear about is dead simple I-want-this-output-from-this-input-optimize-until-you-get-it. That's still impressive: many of those tasks were still impossible ten years ago. But the inspiring stuff is not supervised.
Re: (Score:2)
Fair enough. I did use the qualifier "most", but it's worth noting that there are more complex systems in play as well. Even so, I think the larger point I was making still stands, that unnecessarily high expectations are generated from the term "artificial intelligence", because people naturally define "intelligence" in human terms.
Re: (Score:2)
People get defensive about the term intelligence, because we like to believe we're special. The more you look into it, the less special "human intelligence" seems. Artificial intelligence is an old term, with a reasonable definition:
I don't feel the need to chase the hype ma
Fake it until they stake it (Score:3)
Them, and everybody else (Score:2, Interesting)
Almost anyone touting "AI-powered" anything right now is doing the same thing. The state of the art isn't close to anything resembling genuine artificial intelligence. The kicker is that it doesn't matter most of the time, aside from the marketing department, because AI is not required to do most of the things they are claiming for it.
Re: (Score:2, Insightful)
Another example is voic
Re: (Score:2)
Perfect example: we don't have self driving cars. You just think we do, because of hucksters like Elon Musk. Cars cannot "self-drive" and it isn't clear they will be able to. What "self driving" companies are demonstrating is "lane following" and object avoidance and not doing that particularly well. And no, human intelligence is nothing like the algorithms that computers run. The hucksters use terms like "Neural networks", but that isn't how the brain works at all.
Re: Them, and everybody else (Score:1, Troll)
Re: (Score:2)
Um, humans can do a lot more than lane following and object avoidance. Here is a test to prove what a fraud people like you are: go take a brand new Tesla. Attempt to self-drive it across the country starting at one of Musks mansions in LA, and ending up at the White House in DC. Let me know when you arrive. Complete fraud.
Re: (Score:1)
Re: (Score:1)
Uh, what? You are claiming that humans can't drive from LA to DC? Let me know when you get even out of Musks neighborhood with "self driving". My point is that you are a fraud. You claim all these fantastical things exist, but they don't. You aren't knowledgeable enough to understand the limitations of computer software and hardware. At least Musk is doing it for money, but guys like you live in a fantasy world.
Re: (Score:2)
That's not what I wrote. The fact that you are trying to make it look like that is what I did write shows that even you know how miserably you lost in this very short "debate."
Re: (Score:1)
Right. Here is a simpler test. Go take a Tesla and try to self-drive it across Cincinnati. Until you can do that stop your fraudulent "AI" claims. And by the way, your example of route finding being AI is even more laughable. Those algorithms were developed in the 1950s. So unless you want to claim that AI existed in the 50s you might want to retract that statement too. But I'll see you on Mars really soon now.
Re: (Score:2)
Re: (Score:2)
It is one thing I always noticed about AI/Space Nutters: they never really can prove any of their fantastical ideas actually work in the real world. I guess living in a world where you believe that "space factories" and "asteroid mining" is comforting if you cannot face reality.
Re: (Score:2)
Re: (Score:2)
Perfect example: we don't have self driving cars. You just think we do, because of hucksters like Elon Musk. Cars cannot "self-drive" and it isn't clear they will be able to. What "self driving" companies are demonstrating is "lane following" and object avoidance and not doing that particularly well. And no, human intelligence is nothing like the algorithms that computers run. The hucksters use terms like "Neural networks", but that isn't how the brain works at all.
Ed Balls.
Re: (Score:2)
"Self-driving" is not a binary feature. The NHTSA recognizes five distinct levels of self-driving capabilities [howtogeek.com]. There are already cars with level 1 and level 2 capabilities (driver assistance) out on the road, of course, and a few with level 3 capabilities (conditional automation). Level 4 (high automation) is the tech that's currently being developed. This is a "fair-weather" system, that will still require a drive to help on occasion, and is what's currently expected to deploy in the next decade.
Level
Re: Them, and everybody else (Score:2)
"Node networks" would probably be a better term than "neural networks". That doesn't lead itself to unwarranted comparisons with biology. But it also doesn't sound as cool and futuristic.
I guess it's easier to sell tickets to see the cyborg than tickets to see the well-tuned repetitively iterated algorithm.
Hmmm (Score:5, Interesting)
So they have "advanced technology" to do boring things human technicians used to do to attract billions of dollars of funding, but turns out they're using cheap human techs behind the curtain.
Sounds like another tech startup with wild ideas... how is Elizabeth Holmes getting on these days?
Re: Hmmm (Score:2)
The best comparison would be the Mechanical Turk. There have been many internet-based versions of it from Amazon and others. But (at least Amazon's) didn't straight up lie to you - you knew you were paying a poor person somewhere to send friend requests on FB, or comparison shop for you, or whatever.
I remember a site called "Fiverr" where some of my web hosting clients would get small PHP sites fixed up on the cheap - often from overseas. I doubt what these guys are doing can be that cheap.
Fishing Lures and Toys (Score:5, Insightful)
Know your REAL customers to make the sale.
Doesn't matter if a Fishing Lure catches fish, you just have to hook the Fisherman.
Doesn't matter if kids like a toy, it is adults that buy them. You are really pitiching the item to the adult in the toy store.
Re: (Score:2)
I take it you don't have children and also haven't been around them much. Adults quite often buy toys just to stop the whining; more so now that you can't beat the little bastards. ( <-- I kid. I kid.)
Re: (Score:2)
I take it you don't have children and also haven't been around them much. Adults quite often buy toys just to stop the whining; more so now that you can't beat the little bastards. ( <-- I kid. I kid.)
Every year when the kids were small there would be one present that one of the kids had as their #1 items wanted for xmas that me and the wife would look at and say "they won't play with that more than 30 minutes before getting bored." It wasn't to stop the whining (my kids weren't whiners), but because we knew it was important to the kids that they got their #1 most wished for item ever for xmas. We'd get them the gift knowing full well it would make them happy for a day, and then be forgotten about.
Whin
Re: (Score:2)
Artificial Intelligence (Score:1)
If the intelligence is artificial is it still intelligence?
Lets apply it to some other virtues and see how it sounds.
Artificial Love... no that doesn't sound good.
Artificial Emotion.. hmm no.
Artificial Soul... hmm still no.
They need to AWS the AI link between Azure and Quantum computing, then maybe.
That's not news (Score:3)
We Here See Exaggerated Tech Savy Every Day (Score:2)
But they get paid! (Score:2)
SoftBank-backed startup offers 'human-assisted' artificial-intelligence
So it's a crowd sourcing shop.
The Mechanical Turk Again (Score:1)
> ... the company relies on human engineers in India and elsewhere ...
Looks like a machine, but there are people inside...
Artificial-intelligence technology exaggerated? (Score:1)
Well, a big DOH, first off it isn't Artificial Intelligence, what it is is a pattern recognition engine that under human guidance, may be usefull in extracting answers from a mass of data. AI block-chain basically snakeoil to be sold to the gullable i
I have an AI-enabled bridge to sell you... (Score:2)
Where do I go to get VC funding?