TMTB EOD Wrap: an early taste of HOT AI SUMMER vibes

First, let’s get the war stuff out of the way. Market rallied back to green from down 1.5% early as hopes for a ceasefire and “pushback” of Trump’s ultimatum. All sorts of news reports coming left and right and tough to know what exactly is true, but Axios reported around 12:30pm est that some “progress” had occurred in the last 24 hours and Pakistan’s PM said diplomatic efforts were “progressing steadily".” Here’s CNN after the close:

---

A taste of HOT AI SUMMER vibes…

It’s been a while since we’ve talked about HOT AI Summer vibes here (too long a while, for our taste). The Iran war has everyone navigating in a fog of mixed signals as the factor see-saw continues to frustrate tech investors. Hyperscalers are stuck in ROI capex purgatory. NVDA, AVGO, and AMD have all been rangebound since ASS hit us in the fall. It’s been a long winter for names outside of memory, HDDs, and optical and semi cap (some of which are still at or near 52wk highs). With the dawn of Claude Code and Agentic AI, the vibes have been building but skeptics and lack of full on ai vibes has failed to materialize. A bit of a cognitive disconnect for tech investors.

Here in California we’ve had unseasonably warm weather for the last month - summer came early. Today we finally got an early taste of what a HOT AI SUMMER could feel like for a couple of months if the Iran headwinds were able to lift, thanks to a slew of positive news from Anthropic — first the $30B ARR number in the morning and then more details around their new Mythos model in the afternoon. We are by no means calling an all clear yet, but we feel the fire growing stronger and stronger, and more undeniable underneath the surface of all the very real macro clouds.

Let’s get to it…

More details around Anthropic’s new model Mythos released. Long PDF from Anthropic around capabilities here and here

Here’s TechCrunch:

The model’s limited debut is part of a new security initiative, dubbed Project Glasswing, in which 12 partner organizations will deploy the model for the purposes of “defensive security work” and to secure critical software, Anthropic said. While it was not specifically trained for cybersecurity work, the model will be used to scan both first-party and open source software systems for code vulnerabilities, the company said.

The partner organizations previewing Mythos as part of Project Glasswing include Amazon, Apple, Broadcom, Cisco, CrowdStrike, the Linux Foundation, Microsoft, and Palo Alto Networks. As part of the initiative, these partners will ultimately share what they’ve learned from using the model so that the rest of the tech industry can benefit from it. The preview is not going to be made generally available, Anthropic said, though 40 organizations will gain access to the Mythos preview aside from the partnership.

Mythos Preview can find “tens of thousands of vulnerabilities” that even the most advanced bug hunter would struggle to find. Unlike past models, it can also write the exploits to go with them.
Opus 4.6, the last model Anthropic released to the public, found about 500 zero-days in open-source software — a fraction of Mythos Preview’s output.

Mythos is a general-purpose AI model and was not specifically developed for cybersecurity purposes, Anthropic said. Yet, Mythos has already discovered a number of security issues, Cheng said, including a 27-year-old bug used in critical internet software. The AI system also found a 16-year-old vulnerability in a line of code for popular video game software that automated testing tools had scanned five million times but never detected, Anthropic said

Anthropic is restricting access to 12 launch partners plus 40 additional orgs. That stats look amazing…a real step function up from Opus 4.6, which is already a really good model…

And yes, it’s very expensive:

---

---

INTERNET

• META +35bps: Lots of talk around this article today: Meta Employees Vie for AI ‘Token Legend’ Status from TheInformation:

The rankings, set up by a Meta employee on its intranet using company data, measure how many tokens—the units of data processed by AI models—employees are burning through. Dubbed “Claudeonomics” after the flagship product of AI startup Anthropic, the leaderboard aggregates AI usage from more than 85,000 Meta employees, listing the top 250 power users.

Over a recent 30-day period, total usage on the dashboard topped 60 trillion tokens, according to a copy of it viewed by The Information. The highest-ranked individual user averaged 281 billion tokens, which could cost millions of dollars, depending on the type of model used.

I guess we know who Anthropic’s largest customer is

• AMZN +45bps: hey at least it finished green on all the positive Anthropic news! It’s a win!