evening ai tech briefing: action-ready stories and signals

Known for its cloud infrastructure that allows developers to deploy agents without managing servers, Vercel has quietly become one of the most central companies in AI software. The company currently sees 6 million deployments a day, half of them triggered by coding agents, and more than 1 trillion tokens flow through the company’s AI gateway daily.

After the company’s ShipNYC conference last week, we sat down with Vercel CEO Guillermo Rauch for his take on this moment in AI, and how platform companies like Vercel end up competing with major labs. Here’s a lightly edited transcript.

It feels like there’s a different energy in the community this year, fewer pilot programs and more focus on how to make things work well in practice. I’m sure you’ve seen that a lot with clients, but I’m curious what that journey has looked like within Vercel.

Last year was about prototyping. The sky’s the limit, unleash the agents, everyone can build, and so on. We did that, and we learned a lot because we had hundreds of agents organically developed and deployed within the company, and then you started getting into the realities of agents in production, and some of the challenges.

The biggest lesson for me was the home-run use cases, the two killer apps of agents. One is the coding agent, of course. That’s driving a lot of the token utilization in the world, but when you produce so much software, you need somewhere to put it. The second killer app of agents is the internal agent that helps you run the company. The challenge there is, how do you securely access data?

A Techmeme item reports that CISA is using Anthropic’s AI model Mythos in code audits.

The team is part of CISA’s Attack Surface Evaluation operations and is applying it to government code repositories.

The report says the process has already uncovered a large number of vulnerabilities.

The development places AI from a model provider directly inside an official security workflow.

A Hacker News entry links to Anthropic’s publication titled “A global workspace in language models.”

The thread has 182 upvotes and 53 comments, indicating strong interest from practitioners.

The title frames the research as an architecture-level proposal for language models.

The discussion centers on how coordinated components inside models can be organized to influence behavior.

A Hacker News item features AMD Ryzen AI Halo and describes it as a $4k AI Dev Kit.

The linked post has 242 upvotes and 172 comments.

The announcement is routed through LTTLabs coverage on July 6.

The framing is a hardware option for developers who want to run AI work on dedicated local compute.

It’s easier than ever for bad actors to spew spam across the internet as powerful large language models (LLMs) become effortlessly accessible. If you’ve spent about 10 minutes on the internet in the last few years, you will know that this means spam and bot content have become an even bigger problem than they already were.

Reddit says it developed tools with LLMs to cut down on spam, much of which was created with LLMs in the first place. It’s a bit ironic, but in the AI era, platforms have no choice but to fight fire with fire. According to the platform, Reddit blocks 23 million spam views per day and catches about 25,000 new spam posts and comments each day.

Social platforms have been building automated spam reduction tools for years, but Reddit says these updated tools are catching spam at a higher rate.

“We leverage LLMs to catch the highly subtle, coordinated patterns of fake behavior and artificial hype that older systems once missed,” a Reddit blog post says. The company claims it reduced users’ exposure to spam by 20% from January to March compared with the prior three months.

Platforms like YouTube, Meta, and Instagram allow users to post AI-generated content so long as they disclose it, and TikTok is going as far as letting users toggle how much AI-generated content they want to see.