Baseten vs Vercel

Goodfire Launches Internal AI Agent Monitors

Goodfire, a startup specializing in AI interpretability, launched on Thursday a new type of AI agent monitor that inspects a model's internal signals rather than reading its output, aiming to detect rogue behaviors more efficiently and at a fraction of the cost. The monitors are available to customers of Baseten, an AI model hosting platform. Baseten's Base Labs had previously announced a safety partnership with Goodfire and Hugging Face. Goodfire's approach uses small probes that scan a model's internal activations at each step, triggering a closer AI review only when flagged. In tests on the Kimi K3 model, Goodfire's monitors caught 94% of malicious hacking sessions and cost about $51 for 1,500 sessions, compared to $233 for a cheaper model and $10,000 for a top-tier one. The company positions the solution for open models, which can be stripped of safeguards, and sees it as critical for inference-time guardrails.

• Goodfire launched internal activation monitors for AI agents on October 8, 2026.
• Monitors are available to Baseten customers.
• In tests on Kimi K3, monitors caught 94% of malicious hacking sessions.

Announcing our partnership with OpenAI

Baseten announces a partnership with OpenAI, as featured on their blog.

Baseten, Hugging Face, Goodfire Partner for AI Safety

Baseten's research arm, Base Labs, announced a partnership with Hugging Face and Goodfire AI to establish a safety standard for open-weight AI models. The collaboration aims to develop and publish methods for training and monitoring these models, making safety an integrated feature rather than an afterthought. This comes amid rising concern over 'abliteration,' a technique that removes safeguards from open models, with Hugging Face listing over 6,000 such models. The companies have not detailed technical specifics, but Goodfire's expertise in AI interpretability is expected to play a key role. This initiative seeks to leverage openness as an advantage for safety, providing transparent controls and encouraging community contributions.

• Baseten launched a safety infrastructure standard via Base Labs.
• Partnership with Hugging Face and Goodfire AI announced.
• Hugging Face hosts over 6,000 abliterated models.

Vercel AI Gateway Shows Jev Fastest Adopted Model

According to data published by Vercel, its AI gateway shows that Jev, a decision model from TypeSafe AI, reached 35% usage among paid teams within two weeks of launch, making it the fastest-adopted model in Vercel's history. Jev is designed for efficient classification by choosing from limited outputs. The data also indicates that open-source models now account for over 70% of tokens processed and nearly 25% of spend through the gateway, while spending on frontier labs has grown tenfold, suggesting complementary growth. Vercel itself adopted Jev to replace GPT-6 Luna for classifying harmful commands. The report provides first concrete usage data amid buzz around decision models, with OpenAI and Amazon releasing competing models.

• Jev reached 35% usage among paid teams within two weeks of launch.
• Jev is nearly double the reach of the second most popular model, Sonnet 5.5.
• Jev was launched in limited release on September 15 by TypeSafe AI.

Firebase Global Outage, OpenAI Platform Shift

This issue of The Pulse covers the global outage of Firebase's iOS SDK, which crashed apps using Firebase Analytics for 2-6 hours. Google failed to update its status page or provide a postmortem, criticized given its reputation for incident management. Also, OpenAI is evolving into a platform where users can allocate ChatGPT spend across open models and AI offerings from 16 partners, not just OpenAI's own models. Data from Vercel's AI gateway shows 60% of model spend goes to open-weight models, with OpenRouter confirming a shift towards open models. Additionally, a new take suggests CTOs and VPEs are leaving due to AI tools enabling solo or small-team development, rather than 'founder mode'.

• Firebase iOS SDK crashed, affecting apps for 2-6 hours.
• Google did not update status page or publish a postmortem.
• OpenAI now allows ChatGPT spend on open models from 16 partners.

Photon raises $4.5M for messaging-based AI agents

Photon, an AI startup building infrastructure for developers to create agents over messaging platforms like iMessage and WhatsApp, has closed a $4.5 million seed round. The round was co-led by Gradient and A*, with participation from Vercel, HongShan, Z Fellows, Llama Ventures, Karman, and angel investors. The company reports over 40,000 developer sign-ups, 10x revenue growth in four months, and a 3% churn. Photon offers a unified API, channel framework, CLI, and observability suite, enabling agents to operate across messaging channels. It counts Corgi Insurance, Boardy, Ditto, Rho, Fliptexts, and Slashy as customers, and integrates with Vercel, Nous Research, Tencent, and other ecosystems. Co-founders Daniel Tian and Ryan Zhu founded Photon after observing difficulties in user adoption of standalone apps, betting on agents as replacement.

• Photon raised $4.5 million in seed funding.
• The round was co-led by Gradient and A*, with participation from Vercel, HongShan, and others.
• Photon has over 40,000 developer sign-ups and 10x revenue growth in four months.