SPOTIFY
The podcast host argues we may be experiencing "mass AI psychosis" — a collective overreaction where markets and public perception swing wildly between AI euphoria and scepticism, driven more by hype cycles and promises than verifiable technological reality.
AI companies and markets are trapped in a boom-bust psychology cycle, oscillating between unrealistic optimism when grand projects are announced and sceptical despair when actual capabilities fail to match marketing claims. This pattern resembles the initial dot-com bubble, requiring careful navigation to identify which AI applications will have real staying power versus which represent "AI vaporware."
The Stargate Project exemplifies AI "vaporware flexing" — The $500 billion AI infrastructure project announced with fanfare appears to be facing significant funding and control issues behind the scenes, despite initial market optimism.1 [✓]
Markets react more to AI announcements than actual capabilities — Stock movements for companies like Oracle and IBM have shown volatility based on AI partnership announcements rather than demonstrated technological breakthroughs.2
Claude's COBOL claims triggered market panic without verification — Anthropic's announcement that Claude could modernise COBOL code caused IBM's worst daily drop since 2000 (13.2%), despite limited evidence of actual capability.3 [✓]
Consulting partnerships signal AI's enterprise maturity shift — OpenAI's partnerships with McKinsey, BCG, Accenture and Capgemini represent a strategic move from technology development to enterprise implementation and workflow integration.4 [✓]
The Theranos comparison highlights AI's credibility problem — Like Theranos promising impossible blood-testing technology, AI companies face scepticism about achieving AGI through current large language model approaches, raising questions about "fake it till you make it" incentives.5
General public is transitioning from "wow demo" to "reality check" phase — After three years of impressive demos, users are entering a stage where they realise AI tools often "work nothing like advertised" in practical applications.
"I feel like we're witnessing mass psychosis. It's like you see a demo and think, wow, this is incredible, but haven't made it to the second stage of, oh, this works nothing like advertised."
— Ryan, early in source6"The goal of AI companies is to reach human level intelligence, which is AGI, by the pathway of large language models, which is technically impossible."
— Ewan Morrison tweet cited by host, late in source7
✓ VERIFIED — IBM shares dropped 13.2% on Monday, February 23, 2026, after Anthropic announced Claude Code could modernise COBOL systems. This was IBM's worst daily decline since 2000.8
✓ VERIFIED — The $500 billion Stargate Project was announced in January 2025 by OpenAI, Oracle, and SoftBank at a White House event with President Trump. The project aims to build multiple gigawatt-scale AI data centres across the US.9
✓ VERIFIED — OpenAI announced partnerships with McKinsey, BCG, Accenture, and Capgemini on February 23, 2026, to help sell and implement its Frontier AI agent platform for enterprise customers.10
⚠ UNVERIFIED — The claim that "43% of banking systems" and "93% of ATMs" run on COBOL requires further verification. While COBOL remains critical for legacy financial systems, precise percentages vary across sources and appear exaggerated in the source material.
⚠ UNVERIFIED — The behind-the-scenes details about SoftBank exploring 90% debt financing for Stargate with only $10 billion secured could not be verified through mainstream financial reporting.
For investors: Distinguish between AI announcements and demonstrable capabilities — stock reactions to partnership news may create volatility opportunities rather than reflect fundamental value changes.
For enterprises: Balance AI adoption urgency with verification requirements — consulting partnerships indicate AI is moving from experimental to implementable, but require careful vendor capability assessment.
For developers: Legacy systems like COBOL represent both threat and opportunity — AI tools claiming to modernise legacy code could disrupt established players but also create new career paths in system migration.
The AI industry is entering a credibility consolidation phase where marketing claims must align with practical delivery, creating both scepticism and opportunity for those who can separate signal from noise.
Source credibility: Medium — Podcast host demonstrates industry awareness but lacks technical expertise or direct access to insider information
Claim verifiability: 3 of 5 key claims verified through mainstream financial/business reporting
Potential biases: The host's construction industry background may shape scepticism about tech sector promises; general contrarian stance evident throughout
Quality flags: Inconsistent timestamping; some anecdotal claims without sourcing; mix of verified news and speculative commentary
Confidence in synthesis: Medium — Core market reaction patterns are documented, but psychological framing represents interpretation rather than empirical analysis
Offer: Free trial of no-code/low-code AI platform · Code: Not specified
Category: AI development platform
Credibility: Could not find independent reviews or user feedback about "Area AI" specifically; search results focus on broader low-code/no-code platform comparisons rather than this specific product
Relevance: ✗ Misaligned — While AI platforms align with technological interests, lack of verifiable information about this specific product raises credibility concerns
Ryan, early in source. "Stargate turned out to be a scam... 500 billion in hype a mountain of debt. In reality, the biggest AI vaporware flex in history." ↩
Ryan, early in source. "The market reacted so well to it... these things come in and it just feels like false promises." ↩
Verified via CNBC reporting. IBM shares dropped 13.2% on February 23, 2026, after Anthropic's Claude COBOL announcement. ↩
Verified via Fortune reporting. OpenAI announced partnerships with McKinsey, BCG, Accenture, and Capgemini on February 23, 2026. ↩
Ryan, late in source. "The goal of AI companies is to reach human level intelligence, which is AGI, by the pathway of large language models, which is technically impossible." ↩
Ryan, early in source. "I said that on X yesterday. But that's what it feels like." ↩
Ewan Morrison tweet cited by host, late in source. "They emerged from the same Silicon Valley venture capital startup ecosystem which incentivizes the fast move, the move fast and break things ethos." ↩
Verified via CNBC reporting. "IBM is the latest AI casualty. Shares tank 13% on Anthropic COBOL threat." ↩
Verified via Reuters reporting. "OpenAI, Oracle, SoftBank plan five new AI data centers for $500 billion Stargate project." ↩
Verified via Fortune reporting. "OpenAI partners with McKinsey, BCG, Accenture, and Capgemini to push its Frontier AI agent platform." ↩