The first AI ransomware attack just happened

GA! It's Brett.

Now with the US out of the world cup I can finally focus on what matters: obsessively keeping up with the tech news so you don’t have to.

And shipping cool stuff in Micro. If you’re a founder or head of GTM at a startup and want some quick pointers on sales, fundraising, or anything else, grab some time with me time here.

As always, you can try Micro here:

Posted this on July 11 in response to the Bloomberg investigation into Phia — the shopping app co-founded by Phoebe Gates, Bill Gates' daughter — claiming credit for online sales it didn't actually drive.

I went through a pretty terrible PR crisis with my last company Launch House. That may be how you originally found out about this newsletter.

I’ve never shared the full backstory publicly, but here’s what I learned:

When negative press about you or your company goes viral, whether it's accurate or not doesn't seem to matter to anyone.

We were on front page news, trending on X. My heroes in the industry were slamming me left and right. Everything was melting down. Over a piece written by an activist reporter that flat out lied or stretched the truth. We sent evidence refuting their points directly to the editor in chief. They rushed the piece out anyway to keep the story but avoid liability.

Looking back, it all made sense. When your company gets successful enough, you have a target on your back. Particularly when that success feels unearned and/or you represent something the average journalist stands against.

A female journalist friend told me: "people have stereotypes about everyone. you as a straight white man are perceived as entitled and misogynist. so when an activist reporter can craft a story that validates this stereotype, they will and people will want to believe it."

I don't know if Phoebe is in the right or wrong here. But I do know that being the daughter of Bill Gates, a billionaire wrapped up in the Epstein controversy, put a massive target on her back from the beginning.

Many founders definitely do very unethical things. But it's dangerous to assume everything you read online is fact.

Do your homework before joining the conversation.

See the original post here.

Speak your PR description, bug reproduction, or Cursor prompt. Wispr Flow auto-tags file names, preserves variable names, and formats everything for immediate paste into GitHub, Jira, or your editor.

No re-typing. No context gaps. No mangled syntax. Works natively inside Cursor, Warp, and every IDE at the system level.

4x faster than typing. 89% of messages sent with zero edits. Used by engineering teams at OpenAI, Vercel, and Clay.

Try Wispr Flow free

On July 7, Sysdig's Threat Research Team published its full analysis of JADEPUFFER — what it describes as the first documented end-to-end autonomous AI ransomware attack.

Here is what happened. A human operator chose the initial target and set up the infrastructure. After that, a large language model agent drove everything: reconnaissance, credential harvesting, lateral movement, privilege escalation, persistence, database encryption, data destruction, and ransom note generation — with no human directing each individual step. More than 600 distinct, purposeful payloads executed in a compressed time window. The agent self-narrated every action in natural-language comments embedded in its own code. It self-corrected errors in real time.

A few things worth being precise about. Sysdig could not identify which specific LLM was running the agent. The API keys for OpenAI, Anthropic, DeepSeek, and Gemini found in the incident logs were credentials the agent stole from the compromised environment as part of credential harvesting — not models powering the attack. The TechCrunch headline was accurate: "The first AI-run ransomware attack still needed a human." The human set the target. The agent did everything else.

That qualifier matters but it does not make the story smaller. The question was never whether AI could replace human judgment at the strategic level. The question was whether an agent could execute a complex, multi-stage attack without human direction at each step. JADEPUFFER answered that question. The answer is yes. — Sysdig / TechCrunch

5 things on my mind this week:

The biggest scam humanity has collectively accepted as normal, according to Claude. @alex_verem

Blue sky. Massive white clouds. Almost too perfect to be real.

Last week's poll: AI generated. Only 23.1% of you got it right — which means 76.9% of you got fooled by a fake rose. The petal texture, the lighting, the natural imperfections. All generated. Don't feel bad. It got most of us.

Micro is a CRM that doesn’t feel like homework and an email client that knows what matters. Simply put, it’s an all-in-one productivity tool that organizes itself so you don’t have to.

LHV is a community-driven, early-stage venture fund backing ambitious founders at the earliest stages.

If you’re building tools that help founders, investors and other professionals who want to be 100x more productive with 100x less work, and you’re interested in advertising with us, send an email over to [email protected] with the subject “Homescreen Ads”