Rob's Notes 40: Playing Whack-a-Mole at Infinite Speed
We started a new company...
Note: Not a real JPM Chase Board Meeting about OpenClaw. Source: Google Nano Banana 2
I’ve been busy and haven’t posted in a bit - it’s because we started a new company called InfoHawk earlier this year. So I’ve been talking to people, and building things with my team.
InfoHawk will deliver AI systems and tools to help companies protect their users from scams, fraud and deception.
Why? Well, the Global Anti-Scam Alliance estimates consumers worldwide lost an estimated $442 billion to online scams in 2025. And I suspect this number is still largely driven by non-AI approaches and technologies, for example:
Investment/crypto scams: The FBI seized $15 billion in bitcoin from a “pig butchering” operation in Cambodia in October 2025. Per the FBI, these kinds of fraud complaints are up over 29% over the last year.
Scam ads: Meta alone showed 15 billion impressions per day for scams & banned goods in 2024 (~10% of revenue claimed), per Reuters.
Fake stores and counterfeits: The above figure doesn’t include the global trade in counterfeit goods, estimated at an additional $467 billion per year(!), and various studies over the years point to an enormous number of fake web stores
IBM Research was able to reduce the cost of a successful phishing attack by 96% using ChatGPT in 2023 (slightly lower conversion rate for the AI-made campaign than the human-generated one, but dramatically lower cost to create), but it won’t just be the pros. AI tools will unfortunately further democratize spam and scams, putting their creation within the reach of average people, sometimes accidentally:
“a lot of medium trust systems that relied on there being just enough friction to discourage minor fraud are about to break at scale.”*
There is a lot to work on and we plan to make a dent in these numbers. I’ve been so grateful for the support we’ve gotten from the people in our network already.
Thank you: you know who you are.
This week I saw a security engineering director post to LinkedIn that he was leaving his E8 role at Meta to start a company. What he said feels very much of the moment we are in, and reflects some of my fears as well:
“Much that is solid is about to melt into air in cybersecurity due to AI. We’re staring into a cloud of uncertainty around just how bad things will get as attackers marshal armies of adversarial agents. If we don’t move security professionals from manual laborers to managers of agent teams, we will lose.”
We may not actually get out of having to play whack-a-mole with the worst actors, who are increasingly armed with more and more sophisticated tools and capabilities, and are good at finding gaps in our defenses.
But it just means we have to do more, work together, and we will have to do it all better and faster.
—
Notes:
Counterfeit and pirated trade in the EU
Meta 15 billion scam ads per day
Tweet about medium trust systems
LinkedIn post from security engineering director