Show notes — Episode 2
The week of 12 June 2026. A source under every claim, so you can dig in and check the work.
1. Apple rebuilt Siri on a model it rents from Google
At WWDC in Cupertino on 8 June, Apple introduced the rebuilt Siri: a conversational assistant with its own app that reads the screen and acts across apps, running on a custom version of Google's Gemini in Apple's own data centers. The widely carried figures (a 1.2-trillion-parameter custom model at roughly $1 billion per year) come from reporting, originally Bloomberg in January, not from the keynote itself. The structural point: the company that historically insists on owning every core technology now rents its assistant's intelligence from its biggest rival.
Sources: RedShark News, NPR, TechCrunch (deal background).
2. The two leading AI labs filed to go public
Anthropic submitted a confidential draft S-1 to the SEC on 1 June, after a $65 billion round at a $965 billion valuation. OpenAI followed about a week later; its March round valued it at $852 billion. OpenAI's statement: “We have not decided on timing yet… this gives us the option to go public sooner if that ends up being best.” Until now the AI buildout has been privately funded; public listings mean audited numbers and ordinary investors carrying part of the risk.
Sources: OpenAI, Yahoo Finance, The Tech Portal.
3. The largest IPO in history
SpaceX priced 11 June and began trading on Nasdaq (ticker SPCX) on 12 June, raising about $75 billion at a valuation around $1.75 trillion — roughly 110× its 2025 revenue of $15–16 billion, most of it Starlink (9M+ subscribers). Dual-class shares keep Elon Musk in voting control.
Sources: CNBC, TradingKey. Cut from the draft: a claim that ~30% of public shares were reserved for retail investors — only one source carried it and it could not be verified, so it is out.
4. AI now audits critical software at machine speed (announced 2 June, just before this episode's window; not covered in Ep. 1)
Anthropic expanded Project Glasswing on 2 June from ~50 partners to roughly 200 organizations in 15+ countries, adding power, water, healthcare, and communications. Since the April launch, partners using the restricted Claude Mythos Preview have found 10,000+ high/critical-severity vulnerabilities; of the open-source findings assessed, 90.6% were confirmed valid. A notable example: a wolfSSL flaw (cryptography library on billions of devices) enabling certificate forgery. Anthropic keeps the model unreleased, saying no one has safeguards strong enough yet, and expects other companies to reach Mythos-class capability within 6–12 months, possibly without restrictions.
Sources: Anthropic (expansion), Anthropic (initial update, 22 May), TechCrunch. Cut from the draft: a claim that Mythos found “a 27-year-old bug in OpenBSD” — it appeared in an AI search summary but not on Anthropic's own pages, so it is out.
5. The Iran ceasefire cracked
On Sunday evening 7 June, Iran fired missiles and drones at Israel — its first attack since the April ceasefire — in retaliation for Israeli strikes on the southern suburbs of Beirut. On Monday 8 June, Israel struck the Mahshahr petrochemical complex in southwest Iran and air-defense sites; both sides announced a halt that afternoon (Iran first, Israel at Trump's request). Brent crude rose above $97/barrel. The US-brokered ceasefire (in place since early April, after the war began with US-Israel strikes on 28 February) is meant to reopen the Strait of Hormuz; it has no mechanism for Lebanon, where Israeli operations continue and Iran has said it will intervene.
Sources: Al Jazeera, CNN, Time.