Gotrade News - Anthropic opened Mythos in Claude Code as a public preview, tagged claude-mythos-1-preview across builder surfaces. The launch lands alongside a fundraising round expected near $33 billion.
The preview reignites the US AI capability race as hyperscalers race to monetize agentic workloads. Capital intensity and chip supply will set the pace for AI ecosystem winners.
Key Takeaways
- Anthropic activated Mythos inside Claude Code and Claude Security as claude-mythos-1-preview.
- Anthropic is reportedly in talks to use Microsoft AI chips for next training cycles.
- George Hotz warned AI agents will be costly for software, countering Karpathy.
Mythos Lands In Claude Code
According to Bloomberg Technoz, Mythos appears as an activation button inside Claude Code and Claude Security. The button references the identifier claude-mythos-1-preview for builders.
Anthropic is preparing a broader public release, though tier availability remains unclear. Company materials acknowledge Mythos poses serious global cybersecurity risks, with security framing emphasized.
The rollout sits inside a wider capital push, with a financing round expected around $33 billion equivalent. An OpenAI co-founder is also joining Anthropic in a special role.
Anthropic is reportedly in advanced talks to use Microsoft (MSFT) AI chips for upcoming training cycles. That partnership would tighten an already dense cloud and silicon supply chain.
Karpathy Joins, Hotz Pushes Back
As reported by Bloomberg Technoz, hacker George Hotz published a May 26 post titled The Eternal Sloptember. He said applying AI agents to software will be one of the field's costliest mistakes.
The post landed five days after Andrej Karpathy joined Anthropic's pretraining team. Karpathy cited AI agents as transformative for software development workflows.
Hotz argued that agents cannot code reliably and that the market will take time to realize it. He framed agentic tools as a slot machine lever pulled by hopeful operators.
The split underscores how unsettled agent productivity claims remain. For investors, the question is whether agent revenue scales faster than compute and labor costs.
Stock Read-Through For US AI Ecosystem
The Microsoft chip talks would route more Anthropic training spend into Microsoft (MSFT) infrastructure and silicon stacks. That extends Microsoft's reach beyond its existing OpenAI partnership.
Per Bloomberg Technoz, accelerator demand remains the binding input for frontier training runs. Nvidia (NVDA) stays central as Anthropic, Microsoft, and rivals secure capacity.
Anthropic's existing backers also have direct exposure to Mythos uptake and pricing. Alphabet (GOOGL) and Amazon both hold stakes while running their own frontier model efforts.
The Hotz critique signals enterprise buyers may slow agentic deployments pending clearer reliability data. That risk weighs on revenue assumptions across AI software and tooling lines.
Indonesia's parliament separately flagged national AI governance as too normative this week. The note underlines how policy and funding clarity now shape regional demand for US AI vendors.