AI with Kyle Daily Update 183

OpenAI and Anthropic both moved into consulting on Monday. Same day. By accident? Something in the air!

Anthropic announced a $1.5B joint venture with Blackstone, Goldman Sachs, and Hellman & Friedman. OpenAI announced something they're literally calling The Deployment Company… ~$10B valuation, 19 investors, TPG, Bain, Brookfield, BCG, McKinsey, Accenture, Capgemini, the lot.

Bizarrely both announced on the same day. Same target. They are not selling you software anymore. They are moving in.

Slide deck is below. Today's PDF giveaway is the 2-3 Year Window Playbook - the action steps for doing what these labs are doing, on a smaller scale, in your industry, before they get there.

But first…wtf are they doing here?

What They're Actually Doing

Both ventures send engineers inside businesses. Not "here's the API key, good luck". Engineers walk into real companies, sit with real workers, map real workflows, and rebuild how the business runs around AI.

First argets: healthcare, manufacturing, finance, retail, real estate. The Fortune 500. The PE-owned. Basically BIG BOYS. (That’s where the money is so makes perfect sense).

Funnily enough this is the Palantir playbook. Sit inside. Map everything. Become infrastructure. Once you're that deep, you cannot be ripped out without basically dismantling the company. Palantir is now worth $300B+. Built entirely on this move. It’s a smart one honestly…

The era of "we'll buy some ChatGPT licences and figure this shit out ourselves" is officially over.

If this follows Vampire rules they’ve just been invited in.

The Jagged Frontier Just Moved

Ethan Mollick has been writing about The Jagged Frontier for years now. Hell, I’ve been writing about him writing about it. It’s important though! The idea is simple - AI is brilliant at some tasks, mediocre at others, and the boundary between the two is jagged, not smooth.

For two years that frontier was coding. The reason being… coders built the AI. They knew coding. They optimised for coding. And once you have AI that's really good at coding, you can use it to build better AI. Snowball effect or, if you fancy, “recursive self-improvement”.

Then the frontier moved into writing and marketing - LinkedIn is now a slop avalanche, no offence. Then analysis. And now…

Consulting. Knowledge work. White collar.

Like…all of it.

This is a Trojan horse.

That’s what these new ventures from Anthropic and OpenAI are for. I can’t make this much clearer than this:

• To get into businesses.

• Work out what they do.

• Get AI to do the work.

• Remove the humans.

The Banks Don't Know They're On The List

Here's the funny bit.

Have a look at the list of investors backing this push. Blackstone ($1.3T AUM). Goldman Sachs ($3.7T AUM). TPG. Bain. Apollo. SoftBank. Their logic: AI automates our portfolio and client companies, those companies become more profitable, we win. Bonuses all around.

Cool. But….

Financial analysis. Due diligence. Deal modelling. Research. Reading a thousand-page document and producing a confident verdict. All of that lands directly inside what AI handles well right now.

Translation: the banks think they're running the AI play. They are also on the play. The list. OpenAI and Anthropic are going to come and absolutely each their (rich boy) lunch. Delicious.

Remember: every industry thinks it's the exception. None of them are.

Where We Are

Mental map. Five stages. We're in stage three.

Stage 1 - Coding / software - Donezo
Stage 2 - Content, writing, marketing - Underway
Stage 3 - Consulting / professional services - Now
Stage 4 - Finance, law, accounting - 2-3 years
Stage 5 - Broader white collar - Coming soon.

If your job is sitting in front of a computer moving information around - reports, decks, emails, models, briefs - it's on the list. All of those roles go to zero on a long enough timeline. Five years for most. Ten if you're a bit fortunate. Maybe to retirement if you're really fortunate.

The companies, when they can replace you, will. They don't owe us anything.

Even worse, the ones that don't replace you (out of the goodness of their hearts) will be undercut by the ones that do, by 30-40%, and lose the market regardless. That's the dynamic. Sorry. That's the bit I refuse to sugarcoat!

But.

There's a window before this hits at scale. And that window is where you (personally) live.

What This Means For You

The labs are going after Fortune 500 and PE-owned first. That's where the engineers scale. That's where the money is.

Mid-market is next. Small business after that. There's a 2-3 year window before this reaches your industry at scale. Right now, in the small-and-mid-market layer, demand is huge and supply is basically zero. That gap is yours.

Three moves.

Advisory. Businesses in your sector need someone who actually understands AI and understands the industry. That person barely exists yet. Charge accordingly. Retainer if you can. One-off if not.

Implementation. Don't just tell them what to do. Go in and do it with them. One business transformed properly = referrals, case studies, a track record. Project fees, or a monthly retainer when you build them an internal tool.

Education and content. Become the AI person in your niche before anyone else does. Build the audience. Own the vertical. One of my students, Katie, started doing this for veterinarians. She talks to vets about what they need to know about AI. Vets don't need the Sora updates - who cares. They need to know how to verify AI on diagnoses, or indeed if they should even be using AI at all for that. Helping them with these topics is valuable.

The labs are charging corporations millions for this. You can do it in your industry, at a smaller scale, and still do very very well.

Where To Start

Workshops.

Boring answer. Best answer. Sorry, not sorry!

Get a room of business owners in your sector. Teach them what AI means for their industry and what to do first. Pitch it as basic - because for them, they’ll need basic. If you are reading this now you're light-years ahead, even if it doesn't feel like that (it won't, that's just impostor syndrome talking, ignore it).

Shit just take all my stuff and feed it into ChatGPT to convert it to your niche and go teach it!

My workshop floor is $2,000/hour. Higher rate is $4-5,000/hour. I used to make that in a month - now it’s an hourly rate.

Sounds wild until you do the maths - 400 people in the room for $4000 is $10/person/hour. Cheap as chips for what they get.

If you want to explore this as an option have a look at https://aiwithkyle.com/ai-workshop-series - it’s a free mini-course about how to create your first workshop and market it.

Kyle