Musk is a trillionaire, Anthropic is blocked in India, and Meta admitted it made mistakes. Is AI's golden age already cracking?
Hi,
It’s been a few days now since Fable 5 got banned.
What's your take?
Let’s go.
This week from the AI for Founders podcast on Spotify :
The jiu-jitsu founder who moves slower than everyone and wins — LINK
Jason Katz of Kindling Solutions built a morning AI brief that reads his Slack, his Teams, his calls, and his calendar before he leaves the porch, then interviews him in his own voice so the content that comes out is actually him. The thorn: everyone is an AI consultant now, and single-workflow microservices will cost three dollars on a phone very soon.
The six-person company serving Fortune 500 clients with no human tasks — LINK Elon Salfati of Salfati Group runs a firm where the software literally locks humans out of tasks until an agent decides their judgment is worth the interruption. He explains why agent memory is the next great moat and why most companies are still sprinkling AI to please the board instead of flipping the whole model.
The founder who saw an 'Excel file named Data Lake' and knew everything — LINK Justin Watt of Switchboard studied psychology, not computer science, and that turned out to be his unfair advantage: every business challenge is a human challenge wearing a software costume, and mid-market companies need their workflows mapped before any AI is injected.
The healthcare company cutting front-office costs 50% without firing anyone — LINK
Frederik Mueller and Timm Schneider of Third Way Health pair AI agents with embedded human operators to attack the $350 billion in US healthcare administrative waste, and client Jamie Reddick of Graybill Medical Group cut front-office costs by roughly $3 million over two years without mass layoffs.
Nina (trynina.co) creates SEO-optimized content that ranks on search engines
and gets cited by LLMs like ChatGPT, driving the kind of organic traffic that actually converts.
Investor Corner
Are you raising? Reply here, and we’ll feature your deck.
DiningOut unites restaurants, food lovers, industry suppliers, and brands through a self-reinforcing ecosystem of live events, targeted media, and proprietary technology.
First and only American-manufactured vape cartridge.
Rare, Pristine, Protected drinking water.
The only platform that can authenticate human-made content. Using patented algorithms that analyze how content is made, not just the output.
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.
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.
Voyager is an AI-native Web3 ecosystem and multi-asset exchange. Trade digital assets and commodities in one seamless interface.
RoboReliance is building autonomous robotic systems designed to handle real-world industrial work so humans don’t have to stand next to dangerous, repetitive machines all day.
Tip of the spear:
Anthropic's new models got pulled by US export control — LINK The US government ordered Anthropic to disable access to Fable 5 and Mythos 5 over national security concerns. If your stack depends on frontier Anthropic models, this is a real reminder that geopolitics is now a dependency in your infrastructure stack.
Salesforce just dropped $3.6B to own AI customer service — LINK Salesforce acquired Fin, the AI customer service platform, and will fold it directly into Agentforce. If you're building in the enterprise service layer, your biggest potential acquirer just showed you exactly what they're shopping for.
SpaceX priced the largest IPO in history at $135 a share — LINK SpaceX's IPO is officially the biggest ever, and it made Elon Musk the world's first trillionaire on paper. The signal for founders: the market still has enormous appetite for infrastructure bets with long time horizons.
Elon Musk becomes the world's first trillionaire — LINK SpaceX's IPO pushed Musk past $1 trillion in paper wealth at the same moment tens of thousands of tech workers are being shown the door. That gap is going to produce political and cultural friction that every founder building right now needs to watch.
OpenAI files confidentially for IPO — LINK Hot on the heels of Anthropic's filing, OpenAI has joined the IPO queue. The public market is about to get its first real stress test of whether the AI lab business model actually holds up under scrutiny.
The AI layoff wave is becoming a powder keg — LINK Tens of thousands of workers are being cut while a small cohort of AI insiders is generating wealth at a historically unprecedented scale. The combustibility of that dynamic is real, and founders building workforce tools or agent platforms need to be paying attention to the backlash forming.
Apple's WWDC 2026: Siri gets a major AI overhaul — LINK Apple made its clearest AI push yet at WWDC, doubling down on a deeply integrated Siri across iOS 27 and the broader Apple Intelligence suite. The slow-and-steady approach is looking increasingly like the right one — Apple owns the device layer, and that matters.
Visa just unlocked agent-to-agent payments through OpenAI — LINK Visa integrated its payment infrastructure with ChatGPT agents, letting AI agents execute real commerce on a user's behalf. The agentic economy has a checkout button now.
Framework:
Payroll freeze as AI's primary impact
Most founders are watching the AI jobs story wrong.
The dominant narrative is mass layoffs: AI fires people, humans revolt, regulators respond. But the more precise, more dangerous dynamic is the one happening quietly inside every finance team right now. Companies are not firing people to replace them with AI. They are simply not hiring the next person. The seat never opens. The role never gets posted. The headcount freeze is the layoff, it just doesn't show up in a press release.
For founders building products, this has three concrete consequences.
01. Your buyers are already operating leaner than you think. The person you're selling to has fewer people on their team than they did eighteen months ago. That changes what they need from your product. They are not looking for something that makes their team more efficient. They are looking for something that replaces the teammate they weren't allowed to hire. Position accordingly.
02. Your product should promise a full seat, not a partial assist. The payroll freeze frame means buyers are mentally pricing AI tools against headcount, not against software. A $600/month tool that does the work of a $90,000 hire is not competing with other SaaS products. It is competing with a job req. Build your pricing model and your messaging to reflect that math explicitly.
03. The freeze creates urgency you can use. Companies that chose to automate before committing to headcount are now structurally ahead of competitors who hired first. If you can show a prospect that a competitor in their space already replaced a function with your product, the freeze psychology works in your favor. They don't want to be the company that hired when they could have automated.
The powder keg is real but the explosion isn't coming from mass terminations. It's coming from the compounding realization that the job was never going to come back because it was never posted.
Agentic partnership acquisition for the world’s top podcasts. gethowdi.com
Keep it moving
Meta admitted it made mistakes as AI reshapes 20% of its workforce — LINK Zuckerberg told employees the company made errors during its AI-driven restructuring, which has already touched a fifth of its headcount. When the most data-rich company in consumer tech says the rollout has been bumpy, founders should take notes.
EU orders Meta to open WhatsApp to rival AI assistants — LINK
The European Commission invoked emergency antitrust powers for the first time in 17 years to force Meta to restore free WhatsApp access for competing AI chatbots. Distribution fights are now regulatory fights, watch this space if you're building in Europe.
Claude Fable 5 is Anthropic's most powerful model yet — LINK
The first Mythos-class model leads nearly every benchmark, including SWE-bench Verified at 95%, but comes at double the cost of Opus 4.8 and blocks roughly 9% of requests. More power, more safety filters, higher price, founders building on Claude need to model for all three.
Anthropic walked back its research data policy after backlash — LINK
After researchers complained that a new data retention policy was actively sabotaging their work, Anthropic reversed course. The willingness to walk things back quickly is notable, and a signal that enterprise builders have real leverage when they push back.
Anthropic taps TCS to scale enterprise AI deployments — LINK
TCS is standing up a dedicated business unit to deploy Anthropic's models across its enterprise customers. If you're a Series A company trying to sell into large enterprises, TCS's Anthropic channel is now a go-to-market vector worth understanding.
xAI fired a safety engineer who raised alarms about Grok — LINK
A new lawsuit alleges xAI terminated an engineer for flagging AI safety concerns about Grok days before SpaceX's IPO. The timing matters, this is the story that will follow Musk's trillion-dollar wealth announcement everywhere it goes.
OpenAI wants a 10-gigawatt data center and Nvidia would fund it — LINK
OpenAI is negotiating to lease a planned 10-gigawatt facility in Ohio, potentially financed by Nvidia. The compute arms race is entering territory that only sovereign funds and chip giants can play in.
China is planning a $295B AI data center network to cut out Nvidia — LINK
Beijing is drafting a five-year, roughly $295 billion plan to build a national AI infrastructure network running on 80% homegrown chips. China is not waiting to see how export controls play out.
Google cut its AI subscription price in a warning shot at OpenAI — LINK
Google significantly reduced its budget AI tier pricing, a direct shot at the broader AI subscription market. The race to the bottom in consumer AI pricing compresses margins for everyone building on top of these models.
Mistral is rumored to be raising €3B at a €20B valuation — LINK
The French AI lab is reportedly in talks that would nearly double its previous Series C valuation. Europe's AI horse is still in the race, and the timing alongside OpenAI's IPO filing is not a coincidence.
Coinbase debuted an MCP tool that lets agents trade and pay — LINK
Coinbase launched an MCP server that lets AI agents execute trades and access premium financial data using the x402 protocol. Crypto infrastructure for agentic workflows is no longer a thought experiment.
Ukraine warns future wars will be decided by AI operating systems — LINK Ukraine's defense ministry AI chief says the next phase of war will be a "war of operating systems," with AI networks managing battlefield decisions faster than human commanders can keep up. The implications for AI decision-making at scale extend well beyond defense.
Pokémon Go players' scans are training military drone AI — LINK
Niantic Spatial used roughly 30 billion location scans collected through Pokémon Go to train a visual positioning system now being deployed by defense contractor Vantor in military drones. The data consent story is going to get much louder.
India chills on Starlink right before SpaceX's IPO — LINK
India pumped the brakes on Starlink's expansion plans right before SpaceX went public, a wrinkle that complicates the company's international growth narrative at the worst possible moment.
Oracle hit with a security breach affecting 100+ companies — LINK
A security flaw in Oracle's systems was actively exploited in a mass-hacking campaign affecting more than 100 organizations. If your enterprise customers run Oracle infrastructure, this is a conversation to have proactively.
Google sues Chinese cybercrime ring that used AI to scam hundreds of thousands — LINK
Google is suing the group "Outsider Enterprise" for using AI to run a mass-scam campaign, sending 2.5 million fraudulent text messages over two weeks. AI-powered fraud is scaling the same way AI-powered products are.
China uses DeepSeek while banning Western models — LINK
China is tightening restrictions on Western AI providers at home while US companies increasingly rely on DeepSeek as a cost-effective alternative. The AI Cold War has asymmetric rules.
Dario Amodei has just one direct report — LINK
Anthropic's CEO restructured to have a single direct report. Flat org structures at frontier AI labs are becoming a design choice, not a side effect of early-stage growth.
Amazon turns Alexa into a shopping agent and an ad platform — LINK
Amazon is weaving advertising into AI-powered shopping conversations through Alexa, turning the assistant into both a buying agent and a media product. The monetization model for AI agents is starting to look a lot like the one that won the internet.
DOJ clears the Paramount and Warner Bros. Discovery merger — LINK
The Justice Department cleared Paramount's $110 billion acquisition of Warner Bros. Discovery without conditions. A combined streaming giant changes the content leverage map for everyone downstream.
Snipd is the world’s best podcast streaming app // try premium for free as a sub from this newsletter!
What I’m thinking about:
The real motive behind AI job panic is not what most people are saying out loud. Breakdown in under a minute. — Clip
AI isn't just an engineering problem but a human one. This short reframes what founders actually need to solve for. — Clip
Uncover the AI tools (here) you didn’t ask for, but now you’re weirdly into.
Thanks for being here.
-Ryan
p.s. I write about Revops, Product, and Founder-led marketing on Linkedin , X , and my blog .
