That's not AI, it's being written by 'humans'.

The Builder.ai Incident and the 'Intelligence Fraud' Structure Orchestrated by Masayoshi Son

Rintaro Takechi

Is AI truly 'amazing,' or is it just 'making it look amazing'?
In June 2025, 'LiveCodeBench Pro,' a benchmark testing AI capabilities, was released.
Even with OpenAI's most advanced models, the first-attempt success rate (Pass@1) for difficult problems with an Elo over 3000 is zero percent.

Amidst the flood of discourse claiming that 'AI is surpassing humans,' the actual figures have coldly shattered that illusion.

However, the essential question to ask here is a different one.
Why has AI with such limited performance been 'held up like a god'? And who was orchestrating that illusion?

[Incident] Builder.ai: The Truth Behind the Human-Powered Factory Disguised as AI

The UK-based startup 'Builder.ai' claimed that 'apps can be automatically generated by AI,' and received a valuation of over $1.5 billion from investors including Deepcore (a SoftBank Group subsidiary), Microsoft, and a Qatari sovereign wealth fund.

But its true nature was exposed. The reality was that over 700 engineers were doing the work by hand behind the scenes... That was all there was to it.

This is not a revolution by AI.
It is a story of humans working to the point of death to make it 'look like AI'.

Investors were captivated by demo videos and colorful slides, and not a single one of them looked at the 'substance' that mattered.

This incident is not merely a technical fraud, but a symbol of a structural 'abandonment of judgment'.

[Prophecy] The 'Outsourcing of Intelligence' Depicted in Ghost in the Shell

Many people should feel a sense of déjà vu regarding this structure.
It is from the 2002 anime 'Ghost in the Shell: STAND ALONE COMPLEX,' episode 11, 'The Forest of the Sub-Adults'.

Cyberized children were 'protected' by the state while, behind the scenes, they were handling state-level computational processing. On the surface, it was welfare; in reality, it was the exploitation of intelligence. How is that different from the reality of Builder.ai?

What was there was not that 'AI is amazing,' but a structure that anonymously recycled 'human intelligence'.

[Critique] 'When Will Superintelligence Arrive?': The 'Faith Fair' Hosted by Nikkei

In 2025, Nikkei Visual Data featured a special report titled 'When Will Superintelligence Arrive?', in which 26 experts appeared. And every single one of them said, 'AGI will arrive within 10 years'.

But if you look at it calmly, the perspective that 'it might not come' was excluded. This is no longer 'future forecasting,' but a catalog of 'collective prophecy'.

If capital on the scale of trillions of yen is moving based on such desires, this is an economic sphere that already possesses the structure of a religion.

[Figure] Masayoshi Son: The 'High Priest' of the AGI Myth

The person who has utilized this structure most skillfully is SoftBank Group founder Masayoshi Son.
• 'AI will make all businesses autonomous'
• 'AGI will surpass humanity within 10 years'
• 'I will build the world's largest intelligence platform'

However, what he was selling was not the future.
It was a 'fantasy processed to look like the future'.

The Vision Fund suffered losses one after another, and the sovereign wealth fund concept collapsed.
Even so, the 'Son myth' does not stop.

This is because the narrative itself has been commodified.

[Diagnosis] LiveCodeBench Pro is a device that exposes 'performative intelligence'

In this society of excessive performance, LiveCodeBench Pro is one of the few inspection devices that faces reality head-on.
・Merciless judgment by Pass@1 (if you get it wrong on the first try, you fail)
・Verification using actual competitive programming problems
・Quantitative comparison with humans using Elo ratings
・With non-public scores, there is no room to 'run away' (Google, are you listening?)

It plays the role of a litmus test that exposes the AGI fantasy.
And what it showed was the collapse of that fantasy.

[Final Chapter] The moment you stop asking questions, intelligence dies

It was not AI that was being believed in.
There were only those who deified the 'exterior that looks like AI'.

They get drunk on the performance of slides and rephrase the emptiness of the structure as 'possibility'.
Those who ask questions are excluded for 'not reading the room,' and in that silence, the AI cult sneaks into the core of society.

AI is nothing more than AI. It is an automated information processing device, not intelligence or wisdom.

Masayoshi Son is the one who, while understanding this, shifts the conversation to 'philosophical intelligence' to raise funds.

What he sold was not the future, but a fantasy of a Ponzi scheme that looked like the future.

The moment society stops asking questions, it begins to worship 'intelligence without substance'.
The most convenient thing for that faith is a device called AI that has no accountability.

And even now, someone is silently writing the contents of that AI.

It is your choice to find intelligence in AI.
But what lay beyond worshipping 'fake intelligence' was not a future of wisdom, but a silent black box.

Even so, do you still believe?

Rintaro Takechi