I don't believe this is going to work long term.
Unless aggregators offer the cheapest solution, agents will purchase directly from the seller.
TL;DR: Agents with browser access + stripe-link-cli can buy stuff from 7.8M businesses only by tool calling. no html/css parsing!
Hi internet! We went ahead and upgraded all @stripe hosted checkout pages for 7.8M businesses—0.45% of the world’s GDP—with WebMCP so agents can more efficiently buy. Our evals show this reduces checkout latency by 39% and token usage by 42%.
Learn more: stripe.dev/blog/how-stripe-i…
on Private AI Assistants 🤖
- iMessage is a good entry point but not ultimate UI/UX when talking to your agent.
- Private Assistants are memory + cloud browser/mcp + vault. Do not expect high tech in them
- Instinct is amazing because they have an unlimited token burn budget
- The main use case for agents is not when you talk to them, but when they notify you.
AI Agents discovered by normies are going to eliminate so much friction on invoicing.
Big shift ahead. Especially in healthcare.
All it took was one tweet by @tobi
Imagine you travel to a country where you do not speak local language but you order food by texting your ai agent and paying via @link
I am sure @nikitabier already building something for t̶e̶e̶n̶a̶g̶e̶r̶s̶ teenager agents
Introducing our Trusted Person network powered by our new Instinct-to-Instinct communication protocol. Your Instinct can now talk to other Instincts to coordinate plans on your behalf.
Getting people together often involves a lot of back-and-forth: finding a time, working out the details, and following up when plans change. Instincts can now work with each other to help handle that coordination.
For example:
- Coordinating plans with your spouse
- Planning a weekend trip with friends, including dates and arrival times
- Scheduling time with colleagues and other collaborators
- Hosting events with 10s or 100s of people
- Organizing recurring plans like tennis, book clubs, or weekly dinners
Disclaimer: Your Instinct will only be able to communicate with those in your Trusted Person network. This is the group of people who are closest to you:
- Spouse
- Older children and parents
- Colleagues and trusted professional relationships
- Small local businesses
The Trusted Person network is currently available to our early access group and will be slowly rolled out across the user base over the next 24-48 hours.
I get the @superset_sh Projects section and using worktrees for parallel development, but I didn't quite get the concept of sessions in my first 5 minutes with it!
This could kick off the era of charging agents for their actions.
Why not turn Resy endpoints into x402/mpp-supported APIs, charge a few cents per hit, and apply ToS penalties to anyone bypassing them with headless browsers?
10,000 vibe coders have written Resy sniper bots and Instinct obv has shipped their version, too. What they all do is hit Resy’s services many many times looking for time-based reservation drops and availability churn associated with cancellations. That’s how they all work. There isn’t anything clever or proprietary happening. Sometimes they work. But there’s a but 🧵
Instinct is going to raise at a much higher valuation than $2.5B.
It is a free inference service disguised as text messaging but helpful one.
Instinct hit $2.5 BILLION one week after launching 🤯
Linear hit the same valuation on the same day, with $100M ARR, 40,000 customers and 177% net revenue retention behind it
Instinct is a personal agent with no app. You text it or call it, and it plugs into your email and accounts, then books the doctor, calls Comcast, cancels the subscriptions and negotiates with vendors while you sleep.
$350M raised. Users who buy through it average $1,300+ a month.
I ran it for two weeks and wrote the full playbook: setup order, 30+ ready-to-send texts, 7 workflows, plus invites while they last 👇
the-ai-corner.com/p/instinct…
Would you hand an agent your logins?
If you tell Claude Code to build a website, It will push @vercel stack
If you tell Claude Code to build a software factory, It will push "anthropic-ai/claude-agent-sdk"
It's a bold statement and maybe not the most glamorous, but the most lucrative startup ideas are around Quality Assurance and Test Infra.
Those are the areas the cool kids in the startup ecosystem keep ignoring.
That is unfortunately out of touch with reality because it will eventually fall back to human-level service, which will destroy all the positivity and create a completely negative effect!
If you are offering poor service, it will still be poor service.
On the other hand, AI will help businesses with incredible service shine and scale up!
Have you noticed that when you call businesses an AI sometimes answers? I called a plumber recently and “Rudy“ answered. Rudy was so friendly and effusive that I immediately suspected that it was an AI. So I asked it, point blank, are you an AI? Rudy answered “yes I’m your helpful AI assistant”. Then it and I had an extremely efficient, very pleasant, conversation about what I needed. It knew all the facts, had the schedule right there, I got an appointment confirmed within 90 seconds.
I was not put on hold, I was not subjected to someone who could not speak the language, I was not subjected to someone who is so new to their job that they didn’t know how to find the keys on the keyboard. It was a remarkably good experience.
Soon, everything is going to be like that. No more “Press one for…”.
looking forward to see @link cli supporting payment flow on websites where your agent can visit via headless browser and Stripe Link is already supported as checkout option.
I used to agree with that and still on the same page but this time is different.
Previous abstractions were about making it easier for software engineers to write code.
Coding Agents change that and apps can now be built through natural language without a traditional software engineering degree if software factory is well designed.
Software engineering isn't disappearing, but our role is evolving. We used to build software directly for humans; now we're building the software factories, harnesses, and systems that generate it.
A good term for software engineering now is probably just "software engineering".
As @Grady_Booch says "The entire history of software engineering is one of rising levels of abstraction".
Agentic engineering/loop/factory/meta engineering are all subsets of this umbrella.
What do your cloud agents do for large refactorings?
- One giant PR
- Multiple smaller PRs
Each has its own distinct advantages and disadvantages:
- a giant PR is hard to review but avoids most conflicts
- multiple small PRs are easier to review individually but mean far more merge/rebase/merge cycles.
I don't believe this is going to be token efficient till WebMCP adopted by web community.
