Hey guys!

This was one of those weeks where everything happened at once.

Geopolitics escalated to a level nobody expected.

The AI industry got caught in political crossfire.

Major players shipped products, cut thousands of jobs, and signed military deals, all within days of each other.

Crypto markets whipsawed on war news while prediction markets broke every record and raised serious ethical questions.

A lot happened across many fronts. Here's everything worth knowing from Feb 23 to Mar 1.

Perplexity launched Computer on Feb 25, an autonomous AI agent that orchestrates 19 models (Claude Opus 4.6, GPT-5.2, Gemini 3.1 Pro…) to complete multi-step tasks.

It delegates sub-tasks to specialized sub-agents for research, docs, data, and API calls.

The sub is a bit expensive at $200/month, but honestly for 19 models working in concert with sub-agent delegation, it's not unreasonable.

The press is calling it "OpenClaw for non-techies", cloud-hosted, no self-hosting hassle.

Perplexity CEO explicitly positioned it against OpenClaw. This is the first real consumer product war in the agentic AI space, and it's going to be fun to watch.

Meta's Manus AI added Telegram-based agent controls, operate agents via chat commands for research, inbox management, workflow automation, sub-agent spawning.

If that sounds familiar, it's because OpenClaw pioneered this exact pattern.

They're also testing "Avacado," a new agent integrating both OpenClaw and Manus.

When Big Tech starts copying your architecture, you know you built the right thing.

Trump ordered all federal agencies to "immediately cease" using Anthropic on Feb 27, calling it a "radical Left AI company."

Defense Secretary Hegseth went further and designated Anthropic a supply-chain risk to national security.

The reason? Anthropic refused Pentagon demands about using its AI in autonomous weapons.

Say what you want about the politics, a sitting president blacklisting one of the top 3 AI companies in the world over a safety disagreement is a first. Regardless of which side you're on, this sets a dangerous precedent: build AI with safety guardrails, and the government might punish you for it.

The timing here is almost comical. Within hours of Anthropic getting blacklisted, Sam Altman posted on X that OpenAI had signed a deal to deploy its models on the Pentagon's classified networks.

He praised the DoD's "deep respect for safety." So one company says no to military AI and gets banned. The other says yes and gets the contract. Draw your own conclusions, but the message to the AI industry is loud and clear.

On February 28, the U.S. and Israel launched joint strikes on Iran, hitting security and political targets across Tehran.

Iran's Supreme Leader Ayatollah Khamenei was confirmed dead after his compound was hit.

Tehran launched retaliatory attacks. Bitcoin flash-crashed toward $60K before recovering to ~$66,700.

This is easily the most significant geopolitical event since Russia invaded Ukraine, and it happened on a Saturday, which means equity markets haven't priced any of it in yet. Today is going to be ugly. Whatever your positions are, you might want to check them.

This one's almost poetic. Blockchain investigator ZachXBT teased a major insider trading for Feb 26.

Polymarket immediately created a market asking which company he'd name, $3M+ in volume before the post even dropped.

The target turned out to be Axiom, whose employees allegedly made $200K from insider trading.

But here's the thing: six fresh Polymarket wallets bet $510K on "Axiom" hours before the announcement at 13.8% odds, getting 411% returns.

ZachXBT admitted the leak was "probably inevitable" since he'd contacted Axiom for comment. People literally insider-traded on a market about catching insider traders. You can't write this stuff.

$529M traded on Iran strike contracts. $7B monthly volume (7.5x year-over-year). Record single-day volume of $425M. Company valued at $9B.

But the numbers come with baggage, six fresh accounts made $1.2M predicting Iran strikes, Bloomberg reported meta-insider-trading on the platform's own investigations, and senators are now pushing the CFTC to ban contracts that resolve on someone's death.

Polymarket is simultaneously the most impressive and most ethically complicated platform in crypto right now. The regulatory hammer is coming, it's just a question of how hard.

BTC dropped 5% in minutes when Iran strikes hit, plunging toward $60K before bouncing back to ~$66,700.

The pattern keeps repeating: geopolitical shock → BTC dumps (because it's the only large asset with 24/7 liquidity) → recovery.

Reuters put it bluntly: Bitcoin is "no longer seen as a haven." If you're holding BTC as a hedge against global instability, this week was a reality check.