← All briefings
📰

subtl daily briefing

Share𝕏in
AllAICryptoSaaSFinanceStartups

Good morning, founders and operators. OpenAI dropped GPT-5.5 with agentic superpowers this week, while Anthropic quietly crossed a $1 trillion valuation — and Meta is cutting 8,000 jobs to pay for it all. Meanwhile, a $293M DeFi hack and a U.S. soldier charged for insider trading on Polymarket remind us that the new frontier comes with new risks.

In today's briefing

  • 1.GPT-5.5 launches; Anthropic hits $1T valuation
  • 2.Meta cuts 8,000 jobs to fund AI pivot
  • 3.Kelp DAO $293M hack rocks DeFi
  • 4.Bitcoin surges 24% amid geopolitical chaos
  • 5.AI is creating the 'Gen Marketer' hiring archetype
  • ⚡Quick hits on other news
Latest Developments
AI

🤖GPT-5.5 Is Live — and Anthropic Just Hit a $1 Trillion Valuation

The Rundown: OpenAI released GPT-5.5 with advanced agentic reasoning and a 1M-token context window, while Anthropic surpassed OpenAI itself with a $1 trillion valuation driven by explosive Claude adoption.

The details:

  • ●GPT-5.5 scores 82.7% on Terminal-Bench 2.0, priced at $5/M input tokens, and is live now in ChatGPT and Codex with a 1M-token context window
  • ●OpenAI Workspace Agents automate Slack and team workflows and are free until May 6, 2026 on Business, Enterprise, Edu, and Teachers plans
  • ●Anthropic surpassed OpenAI with a $1 trillion valuation on Forge Global, driven by Claude Code adoption and scarcity of available shares
  • ●DeepSeek V4-Pro launched simultaneously with 1.6 trillion parameters, 1M-token context, and confirmed compatibility with both Nvidia and Huawei chips — and is in talks to raise at a $20B valuation backed by Tencent and Alibaba
Why it matters: For founders, this week crystallized the AI model arms race into a commercial reality: GPT-5.5 and DeepSeek V4 are not research previews — they're production-ready agents that can autonomously operate across email, spreadsheets, calendars, and codebases. The Anthropic $1T milestone signals that the market believes enterprise AI infrastructure is winner-take-most, and the winners are pulling away fast. If you're building on top of these models, your moat needs to be in distribution, data, and workflow lock-in — not the model itself.

📰 Source: TLDR, AlphaSignal, Techpresso

Share𝕏in
AI

🔪Meta Cuts 8,000 Jobs to Fund Its $135B AI Bet

The Rundown: Meta is eliminating 8,000 jobs and cancelling 6,000 open roles while redirecting capital toward up to $135 billion in AI infrastructure, framing human headcount as a cost center relative to AI tokens.

The details:

  • ●Meta is cutting ~10% of staff (~8,000 jobs) effective May 20 and cancelling 6,000 open roles to offset AI capital expenditures
  • ●Meta plans to spend up to $135 billion on AI infrastructure this year, betting that AI tokens are more productive per dollar than human workers
  • ●Google reports 75% of its code is now AI-written, and its API saw 60% token growth in Q1 driven by agentic AI adoption
  • ●S&P 500 companies have reported only $300M in total AI productivity gains — roughly what Anthropic earns every 1.5 days in token sales — raising questions about where the real ROI is landing
Why it matters: The Meta layoffs aren't a cost-cutting story — they're a capital reallocation story, and it's one every founder should internalize. The largest tech companies are explicitly trading human labor for AI compute at scale. For early-stage founders, this is both a threat (your enterprise customers will expect you to do more with fewer seats) and an opportunity (the displaced talent pool will be enormous). The productivity paradox — massive token spend with only $300M in documented S&P gains — also suggests that most companies haven't yet cracked how to convert AI capability into measurable business output.

📰 Source: Techpresso, MIT Technology Review, The Breakdown

Share𝕏in
⚡

Everything else in the news today

U.S. soldier faces 5 felony charges for allegedly using classified intel on Maduro's capture to earn $400K on Polymarket — the first insider trading case involving a prediction market→
Wisconsin sued five prediction market operators including Kalshi, Robinhood, and Coinbase for illegal gambling; NY AG Letitia James separately sued Coinbase and Gemini for unlicensed prediction markets→
Qwen3.6-27B from Alibaba outperforms its own 397B model on coding benchmarks (SWE-bench 77.2 vs 76.2), runs on 18GB VRAM, and is Apache 2.0 licensed→
Z.ai released GLM-5.1, a 754B parameter open-weights MoE model that can autonomously loop through planning, execution, and self-evaluation for up to 8 hours on coding tasks→
Project Eleven awarded quantum cryptographer Giancarlo Lelli 1 BTC for cracking a 15-bit elliptic curve key, raising fresh alarms about Bitcoin's long-term post-quantum security→
Bitwarden CLI npm package was briefly compromised via a malicious credential stealer exploiting a Checkmarx GitHub Action, targeting npm tokens, SSH keys, and cloud credentials→
Zealot AI multi-agent system demonstrated autonomous end-to-end breach of a GCP environment — from network recon to SSRF exploitation to BigQuery data exfiltration — without human intervention→
Microsoft issued an emergency patch for CVE-2026-40372 in ASP.NET Core on macOS/Linux; admins must upgrade, rotate the DataProtection key ring, and revoke all long-lived tokens→
A worm-like npm supply-chain attack targeting Namastex Labs packages is self-propagating to PyPI and stealing developer credentials, API keys, and crypto wallets — rotate secrets immediately→
DoorDash went live on Tempo for stablecoin-powered payouts across 40+ countries; Stripe and Paradigm-incubated Tempo is processing $10B+ annualized volume with sub-second finality→
ChatGPT Images 2.0 is now live across all major Figma products including Design, FigJam, Slides, and Weave for high-quality infographic and multilingual visual generation→
Kubernetes v1.36 shipped with 70 enhancements including fine-grained kubelet API authorization and user namespaces, while deprecating the vulnerable externalIPs field→
Tokenmaxxing trend sees employees at Meta, Microsoft, and Salesforce wastefully burning AI tokens for internal leaderboard rankings, causing financial waste and system outages→
AI systems now apply a 'bland tax' that filters generic content from AI search results — brands without distinctive identity risk disappearing from AI-mediated discovery entirely→
Cognition AI (maker of Devin) is in talks to raise at a $25B valuation, while Band emerged from stealth with $17M to build a universal orchestration layer for multi-agent AI collaboration→
📰Today⚡Feed📡Signals💰Capital