London-based Builder.ai, once valued at $1.5 billion and backed by Microsoft and Qatar’s sovereign wealth fund, has filed for bankruptcy after reports that its “AI-powered” app development platform was actually operated by Indian engineers, said to be around 700 of them, pretending to be artificial intelligence.
The startup, which raised over $445 million from investors including Microsoft and the Qatar Investment Authority, promised to make software development “as easy as ordering pizza” through its AI assistant “Natasha.” However, as per the reports, the company’s technology was largely smoke and mirrors, human developers in India manually wrote code based on customer requests while the company marketed their work as AI-generated output.
↫ The Times of India
I hope those 700 engineers manage to get something out of this, but I doubt it. I wouldn’t be surprised if they were unaware they were part of the “AI” scam.
A billion here, a billion there, after awhile you’re talking about real money.
It’s as if software design is much more than mere coding, and people who need some software written still need a person who can design software, not just code.
So, AI will become yet another tool under a programmer’s belt, much like SQL or Rules Engines (you know, those other tools that were supposed to make programmers obsolete by allowing non-programmer PHBs to create software).
PHBs could learn to design software anytime they wanted, they just don’t want to. They don’t want to think rationally, formulate precise requirements, or think about security, they want to hire nerds to do it for them.
This is btw why I consider all the ThomSeethe™ (that is, Thom’s incessant seething about AI on this website) to be dangerous if you are a programmer. If you don’t start learning generative AI, you’ll soon end up like the programmers who swear by their piles of if-elses and just won’t use a rules engine. Or, in some very pathological cases, even a database.
– or a driver
– or a translator
– or an illustrator
– or a stock-exchange trader
– or a mammographer
– or a travel office clerk
I have many Indian neighbors. But the other day, one had me almost fooled into thinking he was AI. That’s how good Indians have become. When you look at an Indian and they’re indistinguishable from AI, well… it could change the world.
It is depressing that an AI chatbot is considered more valuable than 700 skilled humans.
The next one may sound overly harsh and may get me deleted here for good: In fact I am working in Africa where most software projects are ran by Indian based companies and Indians (for example Infosys). In general Experience is not good and markets are pretty much tired of Indian programmers and software.
Similar experience in Thailand, Indonesia and Sri Lanka although there the sole dependence on Indian software is not as pronounced as it is in Africa.
Rule of thumb: If it is going to be done cheaply, it would always be of poor quality. If the execs charge a hefty price, its still cheap, but exploited labor. There are many Indian devs in the US and even here in SG that are well skilled, but there are plenty more that are not. I’ve even seen this with the big name consulting firms. The same applies with even devs from China.
This is somewhat ironic given how I have talked to people about AI for a couple of years now. I compare it to outsourcing.
At my last company, we made cameras. We had skilled people that used to work for companies that made these kinds of cameras in North America. Today, they are all made in Asia much more cheaply. Typically, we would give the requirements for a design and the foreign design team would turn something around fairly quickly and, again, quite inexpensively. But there were always problems. Maybe the pre-amp would be in the wrong place in the audio circuit causing noise. Or maybe there was a clock mismatch on the board. Or maybe there were security problems in the code. The job of the folks that used to design cameras was to review the design and send it back with change requests. Basically, they were “vibe coding” hardware but with humans. Our software teams were “vibe coding” too. They would ask for stuff from India or Colombia and, when they got it, would find problems and send them back. After some back-and-forth, we had something pretty good. So, the expertise was never really outsourced. But a lot of “the labour” was. This was all very effective and made product development much faster and less expensive than keeping everything in North America. I have seen companies that tried to fully outsource though and it is almost always a disaster. Of course, some places around the world are getting better. Parts of South America have real expertise now. Eastern Europe and much of Asia as well. Those are not typically the companies you “outsource” to though.
Anyway, AI is like a junior employee or an outsourcing exercise. You have to be specific about what you want. You have to know what good would look like if you got it and what is wrong with what you get back if it is not good. You review and provide feedback. Eventually it is pretty good. And, at that point, you have probably saved time and money. But you can only do that if you already have the expertise. If you don’t, you get back the same “not really good enough” first cut and have no idea what changes to ask for. So, more than saving time, you just get very mediocre results. AI is a force multiplier for people that know what they are doing. I can be quite good for that. If you think it replaces those people though, you are wrong.
I mean, maybe someday. We will see.
Given my metaphor of thinking of AI as outsourcing, it is interesting to see somebody literally sub in outsourcing for AI. I bet some people were quite happy with the results.
Fake it till you … file it.
happens to the best ; whoever has not manually updated the SQL database in secret during a demo for a client, let them cast the first stone, although 700 is on another level
Me, quietly dropping the rocks behind me back staring down at the floor. Doing the homer hedge.
Finally, the one instance where saying “AI” in quotes is correct!
boulder.ai == boulder.AllIndians
If you think AI means Artificial Intelligience, then ok.
Actual Intelligence in stead of Artificial Intelligence.
But, how fast were those programmers to scam people? A chat AI usually takes less than a minute to give an answer. Just typing it should take more time.
The indians were probably copy-pasting from ChatGPT.
There seems to be some ambiguity in the reporting of this issue.
I can’t qualify if it was Indian software engineers directly responding to a question, or if it was software engineers being used to vet the questions and responses. The first case is fraud, the second case is quite sensible in the early stages of AI commercialisation.
In fact I’d suggest vetting of questions and replies should be mandatory to eliminate AI hallucinations.
After all AI is a learned state.
“Natasha” is a Russian name. Short for Natalya. Too weird for a simple coincidence.
Do you guys remember last year, when the majority seemed to think ARM was the future of PCs, due to Qualcomm Snapdragon X – remember how there was a minority on here, who knew better. The discourse around AI is starting to feel like that.
CaptainN-,
ARM lacks compatibility on windows,. On the linux side there would be demand except the driver & maintenance situation sucks and it’s not clear that vendors care enough to fix it given that planned obsolescence is seen as a bonus. From a manufacturing perspective, x86 systems that remain viable long past EOL could be deemed the problem rather than ARM.
Honestly I don’t see how any of it relates to AI though. We have a people making the mistake of thinking AI is for them and that AI will fail if they don’t like it, but the long term stability of AI doesn’t stem from end user use cases, it stems from corporate ones.
Even if AI isn’t going to change the world, it definitely changes UI and UX, which is fine by me.
Qualcomm botched the hardware support with poor drivers on both Windows and Linux. The performance did not live up to the hype either, efficiency was worse than Ryzen and M-Series. MS is shooting itself in the foot by tying its ARM fortunes with Qualcomm. Until we see ARM laptop chips from the likes of Mediatek, nVidia and AMD, there won’t be an ARM PC market.
Lovescape offers an ai boyfriend that surprisingly feels like chatting with a real person. I appreciate how it picks up on the vibe and responds accordingly, making conversations flow naturally. The AI even shows empathy at times, which is impressive for a bot. For anyone who needs someone to talk to without judgment or pressure, this is a cool option. Plus, it’s fun to explore the different personalities and settings. Definitely a novel way to experience AI companionship.