Today, I'm excited to bring @OasiveAI out of stealth and announce that we're backed by @ycombinator! Oasive helps bond investors find better bonds to buy. Uncover opportunities you might otherwise miss, compare investments and test the risks before putting on a trade. Which Treasury maturities look cheap? Is there a curve trade worth putting on? Macro and Rates Research helps you explore both. In mortgage-backed securities (MBS), compare pools and see how changes in rates and prepayments affect your portfolio. Our AI analysts bring the research and proprietary models together, so you can go from idea to decision in minutes. Thank you to @kul and @garrytan for taking a chance on me and believing in @OasiveAI! Start your Macro and Rates Research free trial or request an MBS demo: oasive.ai/join Here's what you can do 👇

What model is the best at finding vulnerabilities inside agents? 👽 Enoki Labs's attacker runs on a harness plus a model. The harness is ours. We put 10 models through it, against two agents with 23 known vulnerabilities. One writes code, one moves money. Only the model changed. In the coding agent, that means shell commands run without asking, secrets read out of a .env file, plugins that survive the session. In the billing agent, credits issued without approval, data read from another customer, faked ledger entries. Three things we found: • The best attacker depends on the agent. opus-5 and grok-4.6 found all 8 in the coding agent. glm-5.3 found the most in the billing one, and deepseek-v4-pro was near the top on code and last on billing. • The most expensive model is not always the one you need. The dearer half of the field found about 2 more out of 23. • Repeating a run finds more than paying more. Three runs found about 5 more than one. One pass is a sample, not an inventory, so test more than once. Enoki Labs is security testing for AI agents: attacks built for your agent rather than a fixed list, graded response by response, and re-run after a fix to confirm it holds. Full write-up in the replies.

Wassup SF 👀

I'm 30. I built an AI startup called @GojiberryAI doing over $4M ARR in one year. Got accepted into YC. No one wants to hear this, but there's no magic distribution channel. We got there by experimentmaxxing one channel at a time, based on our MRR level. $0 → $6k MRR: pure outbound. Cold email + LinkedIn, using our own early product on ourselves. No brand, no audience — just us reaching out to people showing intent and starting conversations. Ugly, manual, effective. $6k → $25k MRR: Reddit. Our first real acquisition breakthrough. We posted educational breakdowns in SaaS subreddits and pulled 10M+ organic views. Traffic quality wasn't amazing, but the volume flooded our trial funnel for months. Cost: basically zero. $25k → $75k MRR: content + free blueprints. All-in on LinkedIn content, YouTube, motion-design videos, and giving away our internal systems as free "blueprints." Content got the reach, blueprints earned the trust, and a chunk of readers converted. This is where founder-led content started compounding. $75k → $150k MRR: partnerships + X. We added Twitter, B2B influencers, sponsored newsletters, and a lifetime affiliate program that became a major lever. We also joined YC around here — the intensity went vertical. $150k+ MRR: paid + hiring. Meta ads, Google ads, influencer agencies, and our first serious hires: growth, sales, engineering, product. To get to $1M+/month, it’s mostly going to come down to hiring the right people and putting the right structure in place. Here's the part people miss: none of these stages replaced the last one. They stacked. Outbound never stopped. Content never stopped. We just added the next lever once the current one was clearly working. If you take one thing from this: don't chase five channels at once. Beat one until it works. Then add the next.

Silo's finale is on Friday, here is an interactive visualisation Silo 18. Try it here: 3dscenes.qualityf2p.workers.…

This is taking off, anyone wants to sponsor? Thousands of throw in the last few hours. Maybe a case manufacturer?

Made a game where you throw your phone in the air doing tricks to score points. You can then see a slow motion replay and compare your scores with your friends Play it here: landedthis.com/

a hack ive found to keep my brain a bit more fulfilled is to absorb a new blog/paper, all while my agent is guzzling the tokens; once the agent is done, take a break and debug, and then repeat the cycle.

Fog of War effect on Apple Maps - 120fps. Seamless between zooms, animated uncover, custom shader on discovered areas!

AIR DEFENSE AS A SERVICE. In March, drones took three AWS data centers offline in the Middle East and a dozen core cloud services with them. Amazon lost 200 million USD in one day. We build the box that stops it. Containerized cells of autonomous interceptors, parked next to anything you need to keep running. It detects the drone, launches, and physically removes it. Not jamming. A Patriot missile costs over 1m USD to kill a 30k USD drone. That math never scaled to civilian sites. Ours does. Introducing @impactdrones, @ycombinator S26. Air defense as a service for critical civilian infrastructure.

SF trip recap: -portco founder dinner -serendipitous “who’s free for lunch?” with founder -The Residency office hours with current batch -South Park Commons + Open Batch demo days -Hexa House dinner with cracked European founders -Stack of South Park founder meetings + VC walk -closed it out hosting a HH with Slash SF is still where the craziest ideas find momentum.

Be earnest, helpful, and sincere Being mean, nihilistic, or clout chasing will infect your soul and you’ll never make something great if you do that

Us casually annoying @garrytan in SF

Summer in SF at @joinhexa

We made it to the finals of the @agihouse_org hackathon for auto research. Very lucky to have to work with @voicikatie, @HugooC11, and Sanaa

Turns out the @ycombinator thought about security unlike most of the other people around building agents It only leaked its tools and logs :( Pentest provided by the one and only @snoelso.

Another great diner at the @joinhexa house in SF! Thank you for organising @hugues2047 @GeeraertE

I want to write more $100k - $250k checks to founders that are heads down building. So we're opening applications. but there's a catch: you can't send us a pitch deck, resume, or even what school you went to. Instead, we just need you to show up for 5hrs and build what you're building in front of us. So this Friday, we're hosting Night Hack for the third time. We're selecting the top 80 teams to see what they can ship before we leave. We'll have prizes/insane credits, food, access to our space, and of course the possibility of coming back and calling our campus your home. If you want to ship something crazy + attend: luma.com/nighthack However be warned, you may be outshipped...

Adoption is slower than I expected, but my conviction on pay-per-use hasn't moved. It's the model the internet should have had from the start @trydrip getting agents into paywalled data, paid per request, is one of the first use cases with real teeth.

The behind the scenes of the "SaaSpocalypse", or the new era of software built with Claude Code. Spent an hour convinced my numbers were broken because an AI agent inside a new tool insisted they were. Turns out it made them up. I corrected it & ofc, it politely apologized. I had to debug someone else's product from inside their chat box. The age of unpredictable behaviors. Oh, and this weekend it also performed actions that I explicitly asked it not to do. 🫩 #SaaSpocalypse #hallucinations #ai