It is no secret that I hold the ‘possessors’ of AI at a distance. AI doesn’t exist (not yet at least) and now I got ‘informed’ through Twitter (still refusing to call it X) the following:
So after ‘Microsoft-backed Builder.ai collapsed after finding potentially bogus sales’ we get that the company is entering insolvency proceedings. Yet a mere three days ago TechCrunch gave us “Once worth over $1B, Microsoft-backed Builder.ai is running out of money”, so as such with a giggle on my mind I give you “Can’t have been a very good AI, can it?” So from +$1,000,000,000 to zilch (aka insolvency), how long did that take and where did the money go? So consider this, TechCrunch also gives us “The Microsoft-backed unicorn, which has raised more than $450 million in funding, rose to prominence for its AI-based platform that aimed to simplify the process of building apps and websites. According to the spokesperson, Builder.ai, also known as Engineer.ai Corporation, is appointing an administrator to “manage the company’s affairs.”” Now, I am going on a limb here. Consider that a billion will enable 1,000 programmers to work a year for a million dollars each. So where did the money go? I know that this doesn’t make sense (the 1000 programmers) but to consider that they might accept a deal for $200,000 each, there would be 5 years of designing and programming. Does that make sense? The website Builder.AI (my assumption that this is where they went gives us merely one line “For customer enquiries, please contact customers@builder.ai. For capacity partner enquiries, please contact capacitynetwork@builder.ai.” This is not good as I see it. The Register (at https://www.theregister.com/2025/05/21/builderai_insolvency/) gives us “The collapse of Builder.ai has cast fresh light on AI coding practices, despite the software company blaming its fall from grace on poor historical decision-making. Backed by Microsoft, Qatar’s sovereign wealth fund, and a host of venture capitalists, Britain-based Builder.ai rose rapidly to near-unicorn status as the startup’s valuation approached $1 billion (£740 million). The London company’s business model was to leverage AI tools to allow customers to design and create applications, although the Builder.ai team actually built the apps.”
As such the headline of the Register is pretty much spot on “Builder.ai coded itself into a corner – now it’s bankrupt” You see coding yourself into a corner is not AI, it is people. People code and when you code yourself into a corner the gig is quite literally up. And I can go on all day as there is not AI. There is deeper Machine Language and there are LLM (Large Language Model) and the combination can be awesome and it is part of an actual AI, but it is not AI. As such as Microsoft is believing its own spin (yet again) we can confuse that there is now a setting that Qatar’s sovereign wealth fund, and a host of venture capitalists have pretty much lost their faith in Microsoft and that will have repercussions. It is basically that simple. The first part of resolving this is to acknowledge that there is no AI, there is a clear setting that the power of DML and LLM should not be dismissed as it is really powerful but it is not AI.
As I personally see it, the LLM is setting a stage that the chess computers had in the late 80’s and early 90’s. They basically had every chess game ever played in their memory and that is how the chess computer could foresee what was possible thrown against it. And until 2002 when Chessmaster 9000 was released by Ubisoft, that was what it was and for that time it was awesome. I would never have been able to get as far as I did in chess without that program and I am speculatively seeing that unfold. A setting holding a billion parameters? So I ,might be wrong on this part, but that is what I see and we need to realise that the entire AI setting is spin from greedy salespeople that cannot explain what they are selling (thank god I am not a salesperson). I am technical support and I am customer care and what we see as ‘the hand of a clever person’ is not that, not even close.
So as we are also given “Blue-chip investors poured in cash to the tune of more than $500 million. However, all was not well at the startup. The company was previously known as Engineer.ai, and attracted criticism after The Wall Street Journal revealed in 2019 that the startup used human engineers rather than AI for most of its coding work”, as such (again speculation) a simple trick to replay a mere 1800 days later. And this is what a lot are (plenty of them in a more clever way) but the show is now on Microsoft. They cracked this, so when they come with a “we were lured” or “it is more complex and the concept was looking really good” we should ask them a few hard questions. So whilst we are given “While the failure of startups, even one as high profile as Builder.ai, is not uncommon, the company’s reliance on AI tools to speed coding might give some users pause for thought.” And when we consider “might give some users pause for thought” is a rather nasty setting as I was there already years ago. So where the others? As such we should grill Satya Nadella on “Last month, Microsoft CEO Satya Nadella boasted that 30 percent of the code in some of the tech giant’s repositories was written by AI. As such, an observer cannot help but suspect some passive aggression is occurring here, where a developer has been told that the agent must be used, and so they are going to jolly well do it. After all, Nadella is not one to shy from layoffs.” As such I wonder when the stake holders for Microsoft will consider that the ‘USE BY’ date of Satya Nadella was only good until December 2024. But that is me merely speculating. So I wonder when the media and actual clever people in media are considering that this is a game thatch only be postponed and not won. So will the others run when the going gets tough, or will they hide behind “but everyone agrees on this” as such the individual bond will triumph and there is a lot of work out there. The need to explain to people (read: customers) is that there is a lot of good to be found in the DML and LLM combination. It remains a niche market and it will fill the markets when people cannot afford AI, because that setting will be expensive (when it is ready). These computers will be the things that IBM can afford, as can the larger players like an airline, Ford, LVMH (Louis Vuitton Moët Hennessy) and a few others. But the first 10 years it will remain out of the hands of some, unless they time share (pay per processor second) with anyone who has the option to afford one. That computer will need to work 80%+ of the time to be affordable.
As such we will see a total amount of spin in the coming months, because Microsoft backed the wrong end of that equation and now the fires are coming to their feet. Less then. Less than an hour ago we were given ‘Microsoft Unveils AI Features for Windows 11 Tools’. I have no idea how they can fit this in, but I reckon that the media will avoid asking the questions that matter. As such we will have to wait the unfolding of the people behind builder.ai. I wonder if anyone will ask the specification off what happened to said billion dollars? Can we get a clear list please and where did the hardware end? Or was a mere server rack leased from Microsoft? This is just me having fun at present.
So have a great day and I will sleep like a baby knowing that Microsoft swung and missed the ball by a fair bit. I reckon that this is…. Let’s see there was the Tablet, which they lost against Apple and now Huawei as well. There was the Gaming station, which was totally inferior against Sony. there was Azure (OK, it didn’t fail but a book vendor called Amazon has a much better product, there was the Browser, which is nowhere near as good as Google. And there are a few others, but they slipped my mind. So this is at least number 5, 6 if you count Huawei as a player as well. Not really that good for a company that is valued at 3.34 trillion. So how many failures will we witness until that is gone too?
Have fun out there today.
Lawrence van Rijn - Law Lord to be 