THE BOMBSHELL OF THE WEEK
Amazon announces AI will replace 15% of HR while hiring 250,000 warehouse workers: the first public admission of mass knowledge-worker displacement
Amazon will cut up to 15% of its Human Resources division (the PXT team: more than 10,000 employees globally, which means about 1,500 layoffs already), replacing them with AI. At the same time, the company is hiring 250,000 workers for warehouses and logistics. This is the first time a tech giant has PUBLICLY admitted that automation will massively displace corporate knowledge workers, not warehouse workers.
An AI-equipped recruiter can now do the work of 5 traditional recruiters. Amazon is investing more than $100B in AI this year, but it’s going to cut corporate headcount aggressively.
Its CEO, Andy Jassy, already said in June: “We will need fewer people doing some of the jobs that are being done today. AI is going to reduce our corporate headcount over the next few years.” This wasn’t rhetoric — it’s part of an action plan supporting its strategy and guiding policies.
This is historic. It’s the first explicit admission that AI isn’t coming for the “routine” factory or warehouse jobs, but for the white-collar knowledge jobs previously considered “safe” (recruiting, HR ops, corporate functions). The future of work has already arrived, and it’s not what we expected.
POWER MOVES
Meta lays off 600 AI people while paying more than $300M for Apple executives
Meta is eliminating 600 positions from its AI teams (legacy research, product, infrastructure), but its elite TBD Lab remains untouched. Simultaneously, it’s hiring Ke Yang from Apple, who led AI-powered web search. He’s the third senior AI executive Apple has lost in 2025.
Alexandr Wang (ex-CEO of Scale AI, now Meta’s Chief AI Officer) is eliminating entire teams to concentrate everything in his frontier-models superteam. His memo: “Smaller teams means each person has more impact.”
It’s power consolidation, not cost savings.
OpenAI launches the Atlas browser, Microsoft copies it 48 hours later
OpenAI launched ChatGPT Atlas (a browser with built-in AI) on October 21. Microsoft launched its near-identical clone 2 days later. Coordination between partners, or shameless theft?
OpenAI no longer just makes LLMs — it now controls the entry point to the internet. It competes directly with Google’s Chrome. If they gain traction, Google has an existential problem.
MONEY TALKS
Goldman Sachs acquires Industry Ventures for up to $965M
Goldman is paying $665M in cash + equity with up to $300M more tied to performance through 2030. Industry Ventures has $7B in AUM and an 18% IRR, with stakes in more than 700 VCs.
Goldman is betting big on alternative exits (secondaries, buyouts) because the IPO market remains dead. Industry Ventures reports that tech fund buyouts already account for 25% of ALL liquidity in the venture ecosystem.
Traditional exits (IPO, strategic M&A) aren’t working. Smart money is betting this is permanent.
NUMBERS vs REALITY
Anthropic: the numbers that DO matter
While others inflate metrics, Anthropic reports this:
- $7B revenue run rate (one of the fastest growth stories in B2B SaaS history)
- Claude Code: $500M ARR in 2 months = “fastest-growing product” according to the company
- more than 300,000 companies (vs roughly 1,000 two years ago)
MY TAKE
This week you caught me at TRGCON by David Bonilla Fuertes, so I’m sending this later than usual — but I wasn’t going to miss our date.
Here we go.
The Amazon story is the first domino. It started with code because it moves an obscene amount of money, and because of the very nature of programming (you can test whether code works; you can’t do that with creative writing to see whether a book will be a bestseller).
But needless to say, if they could do it with code, everything else follows. We’ll see how this plays out over the coming months.
I’ll just leave one thought… if they’ve managed to change the paradigm for one of the best-prepared professional groups on the planet… won’t they do it with your job?
Think about it…
The Goldman bet seems clear to me:
Startups no longer go public or get bought by big companies. They’re trapped. Industry Ventures gives them an alternative exit: it buys out the stakes of early investors who don’t want to wait any longer.
My read is that Goldman believes this isn’t temporary — the IPO market died and it’s not coming back. Startups will need new ways to return money to their investors.
On secondaries: traditionally, startups have 2 ways to “exit” and deliver returns to their investors — either an IPO or strategic M&A (Google, Meta, etc. buying you, that sort of thing).
Strategic M&A is getting harder and harder, and startups are getting trapped. You have investors (VCs) who’ve been waiting many years to get their money back, but they can’t get out.
What I think this company does is something similar to what Crescenta, from Eduardo Navarro, does.
They buy the shares of early VCs who want out NOW, even if the startup hasn’t gone public or isn’t going to. The original VC recovers some money, and Industry Ventures waits for the exit.
And the buyouts angle — VERY important.
They buy the whole company with debt, take it private, and keep working it until it’s worth more.
All things considered, I think there are MANY zombie “unicorn” companies that will never be profitable.
As for Anthropic, what can I say — the moment people realize it’s not just for coding but for management too… it’ll be huge.
(It has a “role” feature that can produce, say, a business plan or research done by 50 different people with different perspectives and biases, e.g.)
See you next week!