Co-Founder at Lorikeet. I deleted 10 years of my bad tweets so I could have a public account for my new bad tweets.
San Francisco, CA
Joined February 2010
- Tweets17.9K
- Following2.4K
- Followers1.9K
- Likes22.5K
Pinned Tweet
Personal agents like Instinct and @muse are blowing up. They are flooding companies websites and support channels. This is the B2A (business to agent) era.
Today we're launching Lorikeet B2A: a suite of tools to help businesses win the B2A era.
This is bad analysis:
1. A 1%chance at 79% of all venture proceeds is a great EV!
2. The downside is capped at zero. Asymmetric bets are very valuable.
3. The dataset in the study is undifferentiated. Very easy to believe a particular person’s conditional probability is better than implied by the dataset (easy to be wrong of course too)
Getting VC backing is a fools errand.
New research says 88% of US venture-backed founders receive ZERO financial proceeds from their company ownership at exit, while the top 1% capture 79% of the total exit value.
Not a small study. They looked at 157k startups and 348k founders across the last 40 years. They searched through actual SEC filings and a dozen other sources.
Almost 9 out of 10 got zip. Nada. Zilch. Nothing to show for the stress and even the success of their venture.
Now this doesn’t mean the companies didn’t become profitable or valuable. The VCs know how to structure deals so they get paid a return first. The founders go to the back of the line.
Imagine building a company worth hundreds of millions. Then when it sells you get absolutely none of it. Wow.
Now imagine an alternate universe. You and a cofounder stay away from VC funds. You build businesses that is worth millions or even tens of millions. You can sell it, keep it or close it. If you sell it, you get the proceeds.
I’ve been asked many times why I’ve never taken VC money - it’s never felt like a good deal - and this research shows it’s not… nber.org/papers/w34512?utm_s…
People leaving Ant / OAI after like 14 months and writing War and Peace is a favourite genret of tweet.
Brother you have literally written more words there than you spent minutes working at the company.
I wonder @dkrolph is it possible to be defamed by google’s AI summary of search results?
I know it's a bit inside baseball, but can I take a second to praise Tom and the @lorikeetai marketing team?
Traditional B2B SaaS marketing is all about extraction: create 'lead capture assets' that offer a figleaf of thought leadership in order to push a foregone conclusion and enrol you into a spam email sequence (or worse, a small cold call sequence).
We've taken a different path. We believe the key to marketing in 2026 is to make things your target buyers will find useful and then give them to them for free, in a non extractive way. We have long banned email gates to access our consumer research or other assets. I hate them, and so I won't subject our ecosystem to them.
Take our recently published AX Index. It's a free site (theaxindex.com). You can view the scores for any brand. Any brand can ask us to measure them and we will, and then we'll send the results (without any other gunk). We intentionally didn't foreground Lorikeet in the branding - we felt AX Index will be more useful if it's a piece of ongoing research in collaboration with others in the ecosystem (more on this soon), instead of a sweaty "lead magnet" asset. Nowhere on the site will you see us pushing a Lorikeet product.
Okay so the fun variant of this game is you actually get the math wrong but project enough confidence that the audience don’t catch you. If you get to the answer and they don’t catch you within 5s you win.
Like “2281 times 3 is 6843” you can just lie and see if anyone catches you.
Steve Hind retweeted
What people are failing to understand is that the internet is mostly going away.
We’re not going to be scrolling random shitty websites for much longer, it’s all going to be accessible from our agents of choice.
This doesn’t mean good design is going away, but beautiful websites will be like beautiful coffee shops on a Sunday morning. Relaxing, peaceful, allowing you to take a moment to enjoy life.
Human web designers will always have a place, but it will be an acquired taste for the shit that really matters, not a random government website.
This stuff should just be as functional as possible.
The WHITE HOUSE VIBE CODED their website 😂
And they didn't even hide it
Go to fraud.gov and tell me that is not a raw opus 4.8 output
Every experienced vibe coder I know, knows how to make the output look more natural.
They could have easily used the "make it look less like AI" line - and didn't
We have amateur vibe coders in the white house lol
Steve Hind retweeted
If you want to move money safely with agents, sign up for @naturalpay. It’s what we’re built for
SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios.
The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios.
Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said.
Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior.
Read my latest for Axios here: axios.com/2026/10/22/ai-agen…
There’s a world of difference between user-authorized agents (Muse, Instinct, Grokbot, etc.) and true unauthorized/rogue agents that could constitute cyberattacks.
Most of the activity described in the Axios story appears to be the former: customers sharing credentials with agents so they can take actions on their behalf. This may violate bank terms requiring humans to log in themselves, but these are in no way cyberattacks.
The direction banks should go is to support user-directed agents. For example, Amex is an authorized connection in Muse, making it easy for users to grant, limit, and revoke access.
Read-only permission levels are a sensible interim step. Several banks already support or are moving toward them. Full write access (with user credentials) is a different risk category and should be approached carefully.
Frightening the public by blurring the line between user-authorized agentic behavior and cyberattacks is irresponsible.
I think @BernieSanders has taken the wrong populist, pro-consumer angle here.
It’s BS that big banks block their customers using AI to contact them while they themselves use AI to speak to their customers!
Totally dystopian for a bank to force you to speak of its bot while stopping your bot speaking to them!!!
I think @BernieSanders has taken the wrong populist, pro-consumer angle here.
It’s BS that big banks block their customers using AI to contact them while they themselves use AI to speak to their customers!
Totally dystopian for a bank to force you to speak of its bot while stopping your bot speaking to them!!!
This is insane. We cannot allow AI agents to go into your bank account and act independently without your authorization. But that is what’s happening right now.
Left unchecked, AI could endanger the entire global financial system — including your bank account.
Pause AI NOW!
Readers added context they thought people might want to know
Per Axios, detected AI agents access bank accounts with the customer’s permission. This violates most banks’ terms requiring human logins, which is why banks call it “unauthorized.”
Source: Axios (Oct. 22, 2026) — axios.com/2026/10/22/ai-…
Steve Hind retweeted
there’s a HUGE diff b/w
unauthorized by the bank
vs
unauthorized by the user
if agents are truly acting outside user instructions, that’s scary
but knowing banks, i assume it’s the former so they can fear monger, regulate it away, + kill the risk of deposit flight from agentic money movement
SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios.
The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios.
Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said.
Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior.
Read my latest for Axios here: axios.com/2026/10/22/ai-agen…
Steve Hind retweeted
Muse, Instinct, and Grokbot really changed agentic commerce
From what we can see on @link, total transactions increased by 38x, volume by 20x in September
What are people having their agents buy? Retail was the most popular, then travel, and then music and entertainment
Steve Hind retweeted
M2020, Recap Part 3. Red hot bank issue is Meta Muse in "Agentic Finance". Per previous tweet, consumers not only giving over their banking credentials, but also their OTP. Expect move to structured APIs, explicit consumer permissions that banks hold, data agreements between agentic platforms and banks (with rules and pricing).
Most agree that KYA makes no sense. Banks don't need to know the agent, they need to know what the consumer authorized, and how the consumer was authenticated (by the agent). They also need ensure the agent operator will abide by access terms and rules (ex never asking for OTP).
Steve Hind retweeted
apparently @Muse now at 10.2M US downloads in 40 days per @SensorTower via @TechCrunch (10.9M incl. CAN) --> only the 12th app ever to do that - last one was Threads in ‘23
> DAUs hit 2.5M on Oct 16, up 15% WoW (was 1M+ as of 3 wks ago)
> 30% D7 / 23% D14 retention (vs. 19% / 16% avg for ChatGPT, Claude + Grok)
> ChatGPT had 3.7M US downloads in its first 40 days, Grok 500K, Claude 200K
> #1 on the App Store every day but one since launch
a lot of this is bought --> Meta put 39% of its total ad spend behind Muse in Oct (up from 23% in Sept)
techcrunch.com/2026/10/20/me…
Missed this yday - @Muse is 3 wks old & already has 3M+ people prompting it every week (4M+ weekly users total, 1M+ daily) per @theinformation
- Launched Sept 8, still only live in US + Canada
- 5M US downloads in 22 days (per @SensorTower)
- #1 free app on the US App Store since Sept 18
Distribution, distribution, distribution!
Obvi still super early. Real test is retention + what usage looks like once the initial curiosity wears off and the low hanging fruit use cases (negotiating bills, canceling subs, etc.) get used up
theinformation.com/briefings…
