Episode 57: $965B Anthropic, 1,000 Wix Cuts, and Banks Get GPT-5.5

1 June 2026 · 3:20

Show Notes

• Anthropic raises $65B at a $965B post-money valuation; reports surpassing $47B run-rate revenue in early May
• Anthropic ships Claude Opus 4.8 with 'effort level' controls and parallel subagents; ~4x less likely to miss code flaws; Mythos rolling out in coming weeks under Project Glasswing
• ECB urges euro zone banks to invest more in cybersecurity amid AI risks, referencing Mythos-class models
• Bank of Italy engaging with global AI firms ahead of new model rollouts to address security risks
• OpenAI gives some Japanese financial institutions access to GPT-5.5 to help prevent cyberattacks
• Wix cuts ~1,000 roles (20%); CEO cites the strong shekel and AI reshaping org structure
• IBM / Red Hat commits $5B and 20,000+ engineers to Project Lightwell for open-source security
• EU Startup Fund (EIC) flags up to €4B with potential UK participation for late-stage deep-tech
• AI Tool Spotlight: FairNow for AI governance, bias audit support and NYC Local Law 144 compliance

Key Takeaways

1. The next hiring wave is security, governance and infrastructure — not 'more AI engineers'
2. Anthropic's $965B valuation creates comp and demand pressure for platform, infra/SRE, security and AI governance
3. 'Agent ops' is now a production requirement — evaluation, tool permissions and reliability move into budget
4. Regulators are driving the agenda — ECB and Bank of Italy pull security hiring forward in regulated industries
5. Early-access frontier model partnerships create demand for applied AI security and secure deployment teams
6. Wix's 20% cut adds product and platform engineering supply while validating 'AI reduces layers' management thinking

Frequently Asked Questions

What does Anthropic's $965B valuation mean for hiring?

It signals sustained comp and demand pressure. Expect aggressive hiring in platform engineering, infra/SRE, security engineering, evaluation and AI governance — the roles that keep frontier models and agents safe and operable at scale.

Why do central banks getting involved change AI security hiring?

When the ECB and Bank of Italy start naming AI cyber risk and Mythos-class models specifically, regulated industries move budget forward. That pulls SOC automation, detection engineering, security architecture and model risk and governance roles into 2026 plans rather than 2027.

How should hiring loops adapt to 'more honest' models like Opus 4.8?

Engineers will lean harder on AI-assisted output, so verification discipline becomes the differentiator. Add a 30-minute 'verification station' with a small PR diff or bug report where candidates can use AI tools; score on verification steps, threat modelling and how they handle uncertainty.