TL;DR
Wes Roth argues that AI progress has been a continuous exponential curve all along, not a recent turnaround, and covers Anthropic's new $1.5B joint venture with Blackstone, Goldman Sachs, and Hellman & Friedman to deploy enterprise AI using a Palantir-style forward-deployed engineer model. OpenAI is pursuing a similar $4B venture. The real story is solving the AI deployment gap, not any sudden breakthrough.
Key points
1
Anthropic launched a $1.5B joint venture with Blackstone, Goldman Sachs, and Hellman & Friedman as founding partners, backed by Apollo Global Management, General Atlantic, GIC, Leonard Green, and Suko Capital to deploy enterprise AI at scale.
2
OpenAI is raising $4B from 19 investors against a $10B valuation for a parallel 'development company' targeting manufacturing, healthcare, and finance more broadly.
3
The AI deployment gap exists because implementing AI in real enterprise settings is genuinely hard, requiring both model expertise and deep business domain knowledge simultaneously.
4
Both Anthropic and OpenAI are adopting Palantir's forward-deployed engineer (FDE) model, embedding engineers directly inside client companies to build custom AI harnesses, not just handing off software.
5
Media narratives claiming a recent AI 'turnaround' caused by Claude Code are false — AI capability and revenue have been on an unbroken exponential curve; nothing changed structurally.
Key arguments
→
Extrapolating the straight line on AI capability benchmarks is a more reliable predictive tool than following expert or journalist analysis, per Julian Schritser's argument.
→
The AI tools (Claude Code, Codex, OpenClaw, Cursor) are harnesses around models, not the models themselves — the real competitive moat is the harness plus domain knowledge combination.
→
Companies that get Anthropic or OpenAI systems embedded will become deeply dependent on those labs for updates and maintenance, making this a high-stakes sticky customer acquisition play.
Notable quotes
“Making predictions by extrapolating straight lines on graphs is likely to give you a better model of the future than most experts, even than most actual domain experts.”
“If you can just extrapolate this, if you can just project this in your mind, you don't have to be a genius, just it's a straight line. Can you follow a straight line?”
“There was no turnaround. There was no change. Same world. Just follow the line.”
Worth watching?
⏭️
Worth watching the full video?
The key facts and arguments are all captured here — skip the video unless you want Wes Roth's extended commentary on media bias and the Palantir FDE model explained in conversational depth.
Saved you some time? The creator still deserves a like.
Get summaries like this for your own YouTube channels
Every new video from your channels — summarised and in your inbox before you even knew it dropped.
Try it free