Back to Newsletter
Weekly Dispatch

Anthropic Says It Just Stopped an AI-Assisted Bioweapons Plot, DeepSeek Undercut the Entire Market on Price, and Sam Altman Called His Own IPO Reckless.

September 16, 20269 min read
Anthropic Says It Just Stopped an AI-Assisted Bioweapons Plot, DeepSeek Undercut the Entire Market on Price, and Sam Altman Called His Own IPO Reckless.

Hi,

I’m trying something with a bunch of college kids.

I’m hosting a Zoom where we’re going to build a business from absolutely nothing.

The rules:

It has to be something you can do from home.

• Strictly online.

• No certification required.

• No giant pile of startup money.

And you should be able to get it up and running in about one hour.

Then we’re going to create a content engine that actually attracts customers.

This is for the unemployed, the underemployed, the overwhelmed, the underwhelmed, or anyone staring at their laptop thinking, surely there is another way to make a living.

Probably not for us masochistic founders who live for turning simple ideas into 47-tab spreadsheets and three-year emotional journeys.

If this sounds like you, join us.

If it sounds like somebody you know, forward this to them.

We’re going to see how far you can get from “I need an idea” to “I have a business” in one hour.

Reply “LETS GO” and I’ll send you the invite

(or the video if you can’t make it to the live session.)

Let’s go.

Free competitor analysis from CEOfriend

This week from the AI for Founders podcast on Spotify:

The $100M Founder Who Fired His Own Advisory Board and Built a Better One — LINK

Steve Phillips built Zappi to roughly $100M ARR, then stepped back to Executive Chair and turned the advisor relationship into software. CEOfriend.ai gives founders a CFO, entrepreneur, and investor AI advisor on call, starting at $29/user a month.

This Startup Trains Dogs to Sniff Out Four Cancers From One Breath — LINK

Shlomi Madar's SpotitEarly turned a dog's nose into an auditable lab result, hitting 93.9% sensitivity across four cancers in a 1,386-person double-blind study. It's built as a triage layer, not a replacement for your mammogram.

The Recruiting Critic Who Built an AI That Replaces Recruiters — LINK

Shar Broumand's Zero Hiring runs the whole funnel: it interviews the founder, drafts the JD, screens candidates with a 15-minute AI interview, and only charges 6% of salary when someone's actually hired.

Skillz of the Week

Job Interview Loop Designer

Takes any open role and builds a structured multi-interview hiring loop where every interview has exactly one assigned job: competence, problem-solving, resume verification, or culture and values. It writes the scorecard for each stage, assigns who runs it, and outputs a champion-or-veto decision protocol modeled on the FreeMarkets round table Shar describes in this episode.

When to use it:

Any time you are about to open a role and your current process is “a few of us talk to them and compare notes,” or when you need to diagnose a loop that keeps producing hires who interview beautifully and perform badly.

Benefits:

You stop letting one strong impression contaminate every evaluation. The founder who falls in love with people gets a seat where that instinct is an asset instead of a liability, and you get a written artifact you can hand to an AI recruiter, an agency, or your first talent hire without re-explaining your standards every time. Save the download as interview-loop-designer/SKILL.md in your Claude or Chat skills directory, then bring it to your next open role.

Click to steal this skill

AI + Human Stylists → LEARN MORE → Save $30 with RYAN30

Tip of the spear:

Anthropic

Dario Amodei laid out a plan to pace Anthropic's frontier releases deliberately rather than race flat out. For founders building on Claude, it's a signal that model cadence itself is becoming a strategic lever, not just an engineering constraint.

Anthropic Caught Someone Trying to Weaponize Claude — LINK

Anthropic says it detected and blocked an attempt to use its models to support biological weapons development. It's the clearest public proof yet that frontier labs are actively fighting the worst-case misuse scenarios, not just talking about them.

DeepSeek Just Made Frontier Models Feel Overpriced — LINK

DeepSeek V4.1 Flash landed with a $0.003-per-million off-peak cached input rate and benchmarks that beat GPT-5.6, Sol, and Claude Opus 5. If the numbers hold up, every lab's pricing page just got a lot harder to defend.

Apple Finally Made Its Hardware Event Weird Again — LINK

Apple's fall event delivered a foldable iPhone Duo and an always-listening Apple Watch, its biggest hardware swing in years. For founders watching where always-on ambient AI actually ships first, this is the tell.

Miro Just Sold for 90% Less Than Its 2022 Price — LINK

Bending Spoons is buying Miro for $1.36 billion, a 90% haircut from its 2022 valuation. It's a blunt reminder that collaboration tools without an AI-native wedge are getting repriced hard, whatever they raised at during the boom.

AI Coding Just Proved It's Not Winner-Take-All — LINK

Cognition hit a $48 billion valuation, and investors are betting that AI coding has room for more than one giant winner. Good news if you're building an AI dev tool and dreading a Cursor-shaped ceiling.

Sam Altman Says an OpenAI IPO Would Be a Mistake Right Now — LINK

Altman called it ill-advised for OpenAI to go public in 2026, even as pressure builds from investors and compute costs. Read as a signal that the math on public-market scrutiny still doesn't pencil out for them yet.

150 Million Driver's Licenses Just Leaked — LINK

ID-verification giant IDScan confirmed a breach exposing more than 150 million driver's licenses, one of the largest identity data leaks on record. If your product touches KYC or age verification, assume your users are asking about this.

Investor Corner

Are you raising? Reply here, and we’ll feature your deck.

Jobs2rely.com

An autonomous AI job search agent for senior tech professionals. It runs a full pipeline: discovers relevant jobs, scores them against your profile, rewrites user's resume authentically for each role (not generically), finds the right hiring manager, and drafts outreach

Wayzn

Other pet doors are ugly, poorly insulated, and can be difficult to install. Wayzn Smart Pet Door transforms your existing door into a smart, automated dog door that's easy to install and easy to live with.

Dining Out

DiningOut unites restaurants, food lovers, industry suppliers, and brands through a self-reinforcing ecosystem of live events, targeted media, and proprietary technology.

Trinity Water

Rare, Pristine, Protected drinking water.

Human Intelligence®

The only platform that can authenticate human-made content. Using patented algorithms that analyze how content is made, not just the output.

Cora

Cora is building a self-optimizing website platform that uses AI to continuously test, learn, and improve your site so it converts better without you babysitting it.

EmpathEQ

AI-powered training simulations to help healthcare learners practice complex interpersonal and communication skills. We focus on realism, repeatability, and scalability—making soft-skills training more accessible and measurable.

Subscribe and save 35% here.

Keep it moving

Anthropic Says Alibaba, Moonshot, and DeepSeek Have All Been Copying Its Models — LINK

Anthropic detailed distillation campaigns it traced back to three major Chinese labs training on Claude's outputs. Expect this to keep shaping the open-weight-versus-closed debate all year.

OpenAI's Fight With Mathematicians Isn't Cooling Off — LINK

The public spat between OpenAI and mathematicians over benchmark claims kept escalating this week. Worth watching if you cite AI benchmark results in your own pitch deck.

OpenAI Paused New Pro Signups Because Demand Broke the System — LINK

OpenAI put new ChatGPT Pro subscriptions on hold after demand for its Astra product outpaced capacity. A good problem to have, but a rough look for anyone paying for Pro access right now.

OpenAI Wants to Build Your Investor Deck for You — LINK

ChatGPT for Financial Services pulls research, cites sources, and builds decks in minutes. If you've ever paid an analyst to do exactly that, this is the product to watch.

Revolut Got Fooled by Fake Government Data Requests — LINK

Revolut confirmed a customer data breach after attackers impersonated government requests to extract information. A reminder that your verification process is only as strong as the request you didn't double-check.

Meta's AI Agent Is Now Bigger Than Almost Everyone Else's — LINK

Muse climbed to the No. 2 spot among US apps, proof that distribution still beats a better model when Meta's the one shipping. Consumer AI is now Meta's game to lose.

Done-for-you brand sponsorships for podcast creators.

Automattic's Board Just Sidelined Its Own Founder — LINK

Automattic's board forced Matt Mullenweg into a leave of absence, an unusually blunt move against a founder-CEO. Worth a look if you're a founder wondering how much runway your own board patience actually has.

Tesla Finally Picked a Date for the Roadster — LINK

After years of delays, Tesla says the second-generation Roadster gets unveiled October 1. Believe the reveal when you see it ship, not when you see the date.

Robots Need Training Data Too, and Sequoia's Betting Big on Who Supplies It — LINK

Mecka AI is nearing a $500 million valuation in a Sequoia-led round, riding the rush to supply training data for robotics. The data layer for physical AI is turning into its own gold rush.

This Startup Walked Away From $1.5 Billion to Talk to Salesforce Instead — LINK

AI research startup Listen Labs scrubbed a $1.5 billion funding round to pursue talks with Salesforce instead. When an acquisition offer beats your term sheet, that tells you something about where the real leverage sits.

Y Combinator's Garry Tan Wants US Labs to Start Copying Too — LINK

Tan argued US open-weight labs should distill frontier models the same way Chinese labs have. A notable stance from someone who funds a lot of the labs this would affect.

Kimi's Maker Is Targeting $2 Billion in Revenue — LINK

Moonshot AI, maker of the Kimi models, is targeting $2 billion in annual revenue. Another data point that open-weight Chinese labs are scaling revenue, not just benchmarks.

Create a new corporate Uber account here and receive up to 10% off. Other terms and conditions apply, see full terms.

What I’m thinking about:

Is your AI chatbot legally liable when a customer falls in love with it? — Clip

Thanks for being here.

-Ryan

p.s. I write about Revops, Product, and Founder-led marketing on Linkedin , X , and my blog .

Written by

Ryan Estes
Ryan Estes

Investor • Founder • Creator

Ryan Estes created AI for Founders, a podcast, newsletter, and workshop platform reaching 47,000+ entrepreneurs and CEOs. Based in Denver, Colorado.

More Issues