AI in a Nutshell - February 2026 - The Pentagon vs Anthropic, Jack Dorsey Fired Half His Staff, and Software Stocks Lost a Trillion Dollars
In other news... Anthropic tells Pentagon “no”, gets blacklisted, and climbs to #2 on the App Store. OpenAI closes the biggest funding round in tech history. Block fires 4,000 people and blames AI.
Fellow human, it’s the end of February. AI hasn’t taken over the world yet. But it tried really hard this past month.
This is the first monthly edition of the newsletter. And to be honest, I was looking forward to this one. I’ve been away for a big part of this new year, and it seems like for every hour I sleep, there’s something new happening in the world of AI, and I’ve not really had time to sit down and run through the news. So here I am. And by the way, this one is a long read, not sure I can particular tag it “in a nutshell”, but I’ll try 😃
What You Must Know
Anthropic Told the Pentagon “No” and Got Blacklisted → This was the biggest story of the month, and maybe the biggest AI story of the year so far. The Pentagon demanded that Anthropic remove safety guardrails from Claude for military use, specifically for autonomous weapons and mass surveillance of US citizens. Anthropic refused. Trump got pissed and ordered all federal agencies to immediately stop using Anthropic. Trouble galore.
OpenAI Signed the Same Pentagon Deal… with the Same Safety Rules → Hours after the Anthropic ban, Sam Altman announced OpenAI had secured a deal with the Pentagon for classified military systems. The confusing part was that the agreement reportedly includes the exact same guardrails Anthropic insisted on: no mass surveillance, human control over weapons. The internet had a field day with this. Dario got a blog post. Sam got the contract. People bashing OpenAI on Twitter. Interesting times 😎
Claude Hit #2 on the App Store After Getting Banned → In possibly the greatest accidental marketing campaign of 2026, Claude surged to #2 on the App Store overnight after the supposed ban. Turns out, people now trust Claude more because of their honesty, and people now hate OpenAI more. At least, based on what Twitter is saying.
OpenAI Closed the Biggest Private Funding Round in Tech History ($110 Billion) → I had to read this number multiple times. To put it in perspective, this single round is larger than the GDP of most countries. No wonder they needed that Pentagon deal 😄
Claude Cowork Plugins Triggered a $285B Stock Crash (the “SaaSpocalypse”) → Back at the start of the month, Anthropic released 11 plugins for Claude Cowork, and software stocks lost $285 billion in a single day. The term “SaaSpocalypse” trended all month. The fear is that AI agents will replace entire SaaS products rather than just assist them. Companies like Salesforce and Intuit took double-digit hits. It’s early days, but Wall Street appears to be taking the threat seriously.
Jack Dorsey Fired ~4,000 People and Said AI Did It → Block (the company behind Cash App) slashed roughly 40% of its staff, and Jack Dorsey didn’t sugarcoat the reason. He posted a memo on X explicitly saying AI has “fundamentally changed what it means to build and run a company.” Their stock went up 17%. If this doesn’t signal where things are heading, I don’t know what does. And my concern is, he just set the pace for other companies to follow.
Karpathy Said “The Programming Era is Over → Andrej Karpathy (ex-Tesla AI, OpenAI co-founder) posted a thread describing how he built a full video analysis dashboard by giving AI agents instructions in plain English. It was done in 30 minutes with zero manual coding. His exact words: “You’re not typing code anymore, you’re spinning up AI agents, giving them tasks in English, and managing their work in parallel.” He called it “agentic engineering.” I know that whether I agree with the framing or not, the trend is undeniable.
Anthropic Raised $30B at a $380B Valuation → Despite everything that happened later in the month, Anthropic closed a $30 billion fundraise at a $380 billion valuation, making it one of the most valuable private companies ever. It’s interesting that the same company that was blacklisted by its own government is also one of the most-funded AI companies on the planet.
Perplexity Launched “Computer” → Perplexity dropped a new AI product called “Computer” that can browse the web, complete multi-step tasks, and interact with websites on your behalf. From my use of it, it sits pretty close to Claude Cowork and Openclaw.
What’s Good to Know
Samsung Galaxy S26 Ships with Gemini Agents → Samsung launched the Galaxy S26 with (as they claimed) deeply embedded Google Gemini agents that can manage settings, schedule tasks, and control third-party apps. I’m not really motivated, considering how sloppy Siri or Apple Intelligence has been within iOS.
Anthropic Unveiled Enterprise Agent Integrations → Claude Cowork now integrates with Slack, Intuit, DocuSign, and Gmail. The framing was that AI will augment enterprise tools, not replace them. Twitter peeps didn’t fully buy it, but it partly helped software stocks bounce back.
Google Released Nano Banana 2 → Google dropped a lightweight image generation model designed for on-device use. The idea is real-time AI image creation on your phone without needing the cloud.
A Study Says AI is Making Work Harder, Not Easier → Some Harvard Business Review study claimed that AI tools are actually making knowledge workers busier. Employees report spending more time managing and verifying AI outputs than on the tasks AI was meant to automate. I beg to differ. I agree that you’ll spend more time rambling around with AI if you don’t know how to use it, but that’s just a human skill issue, not an AI issue, in my opinion.
Anthropic Acquired Vercept → Anthropic bought Vercept, a startup focused on AI-powered document understanding, signalling a deeper push into enterprise automation.
Kaizen Corner → Just my honest opinion
I’ve been writing this newsletter for 7 months now, and February 2026 was the first month when everything felt… different.
And I’m not saying this because the models got smarter (they did, but that’s normal now). But the consequences are beginning to look more real.
The repeated pattern I used to joke about (“new model, China did something, Europe regulates, AI Godfather warns us”) is gradually being replaced by something much harder to summarise.
I don’t really have a neat way to explain it. I just know that when I sat down to write this edition, it felt deeper than usual. And I spent so much time going down the rabbit hole. Because I need to really understand this AI thing, I don’t know about you.
Nevertheless, stay curious and stay employable 😎
Here’s a gem I found that I think can be helpful → https://anthropic.skilljar.com/
Meme of the Month
That’s your month in AI.
If you learned something, tell a friend. And if you didn’t, blame yourself :)
Until next month,
Kay - your fellow human.
P.S. If this email lands in spam, that’s your inbox trying to stop you from staying plugged in. Fix it.
Quick Note: I’m not an expert, I’m just a curious normie navigating the world of AI (like you). I used to do this every Sunday morning before the kid wakes up. Now it’s once a month, but the ritual is the same: I go online, read about AI, and write down everything I found for anyone curious enough to read it. Why? Because I learn faster and understand deeper when I have to put it in writing. You’re welcome.