- TheTip.AI - AI for Business Newsletter
- Posts
- The AI is building its own successor
The AI is building its own successor
26% of Anthropic's R&D is now done by Claude 👀
Jeff J Hunter
September 18, 2026
The Machine Builds the Machine
Anthropic just published a number nobody's ever published before.
26% of its research and development is now done by Claude.
Not assisted. Done.
A quarter of the work that builds the next Claude is being performed by the current Claude.
That's recursive self-improvement.
The thing safety researchers have warned about for years, where AI helps build better AI, which helps build better AI.
And Anthropic didn't hide it.
They put out three measurements to track it publicly.
How much AI R&D is done by AI.
How much agent activity has a human watching it.
And where the compute is going.
Their words...
AI systems are increasingly being used to build the next version of themselves, and they want to illuminate that progress for the public.
Remember Amodei's essay from Monday?
He specifically flagged recursive self-improvement as the reason to pace the frontier.
Four days later, his own company measured it and showed its work.
That's exactly the transparency the essay asked for.
Now here's the part that lands in YOUR hands.
Claude Code just rolled out something called Projects.
You give it one conversation.
It splits the work into threads on its own.
Runs them as parallel sessions in the cloud.
Passes context between them.
And keeps going after you close the laptop.
In beta now for select Pro and Max users.
So put the two stories together.
The AI is building a quarter of its own successor.
And the version in your hands is learning to run multiple jobs at once without you present.
That's not a future headline.
That's this week.
Here's what I want you to take from it.
The gap between "AI is a tool I use" and "AI is a workforce I manage" is closing fast.
Anthropic already crossed it. A quarter of their engineering is agent work.
The operators who learn to manage that workforce, with real scopes and real oversight, are the ones who compound.
Everyone else is still typing one prompt at a time.
Learn to run the team. 🫡
Did You Know?
"Bootstrapping" comes from the old phrase "pull yourself up by your bootstraps"...
It was originally a joke about something physically impossible.
Then computer scientists borrowed it for a machine that loads its own operating system on startup.
Seventy years later, the machine isn't just starting itself.
It's building the next one. 🥾
In partnership with
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.
LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.
The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.
Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
🗞️ The Tip AI News 🗞️
The public wants brakes on a car most businesses haven't started.
A new Politico poll landed this week.
A majority of Americans now believe continued AI development will destroy humanity.
A plurality want to pause it entirely.
Geoffrey Hinton, the "Godfather of AI," was on Capitol Hill yesterday warning Congress.
And an Anthropic researcher's resignation post, warning superintelligence could kill us all by decade's end, has racked up 800,000 likes.
The fear is real, it's mainstream, and it's bipartisan.
Then Fortune published the other half of the picture.
Corporate America is moving too SLOW on AI.
Not too fast. Too slow.
Most companies are still stuck in pilots, cautious deployments, and hiring plans that haven't caught up.
Some of it is genuine caution.
A lot of it is that they simply don't know how, and can't find people who do.
So here's the strange shape of the moment.
The public is terrified of an AI future.
The labs are debating whether to slow it down.
And the average business hasn't even left the driveway.
That gap is the entire opportunity.
The companies that will actually adopt AI in the next two years aren't waiting for the safety debate to resolve.
They're waiting for someone to show them how to do it carefully.
Someone who takes the risk seriously AND knows how to build.
The scared won't do it. The reckless will do it badly.
The careful operator in the middle gets paid to do it right. 🫡
😂Meme of the Day
Over to You...
Go learn to run the team this weekend, not just the prompt...
The machine's already managing itself. 🔥
Founder, AI Persona Method | TheTip.ai
PS. Full DC recap coming Monday... The Summit, the talk, and how the Gala went. 👀
Get paid to solve problems like this → AI Certified Consultant
 | » Join the AI Money Group « 🚀 Zero to Product Masterclass - Watch us build a sellable AI product LIVE, then do it yourself 📞 Monthly Group Calls - Live training, Q&A, and strategy sessions with Jeff |
Sent to: {{email}} Jeff J Hunter, 3220 W Monte Vista Ave #105, Turlock, Don't want future emails? |
Reply