AI Projects - October 16, 2026

Week of: October 16, 2026

---

Overview

The week’s AI projects news was led by major business signals from Anthropic and Modal, alongside applied case studies in biotech, financial agents, and marketing automation. Practical guides for Claude integrations and AI search tracking also stood out for teams building and measuring AI-powered workflows.

Stories

1. Anthropic tells investors quarterly revenue up nearly 50% to $16.8bn ahead of IPO

Source: Simon Willison Link: https://simonwillison.net/2026/Oct/15/anthropic-tells-investors/

Anthropic told investors that its quarterly revenue rose nearly 50% to $16.8 billion ahead of its IPO, according to the article. The disclosure places the company’s pre-IPO financial trajectory in focus.

For AI builders, Anthropic’s revenue scale signals sustained enterprise demand for frontier models and agentic systems. It also reinforces how major model providers are becoming large businesses whose pricing, product, and platform decisions will shape the tooling ecosystem.

Impact Analysis: Frontier model providers are now major revenue businesses, so builders should expect enterprise-oriented platform shifts ahead of IPO cycles.

2. Modal’s $1B ARR, $16B Series D makes it NYC’s newest AI Infra Decacorn

Source: Latent Space Link: https://www.latent.space/p/ainews-modals-1b-arr-16b-series-d

Modal reached $1 billion in ARR and raised a $16 billion Series D, making it New York City’s newest AI infrastructure decacorn, according to Latent Space. The report cites 3x usage and 3x valuation growth in five months, and 15x growth in one year, with the snippet noting “sandboxes are all you need.”

This matters for AI builders because Modal provides infrastructure for running sandboxed workloads that support agents and AI applications. The scale of its growth highlights how much demand there is for compute environments purpose-built for production AI projects.

Impact Analysis: Sandboxed AI infrastructure is becoming a critical dependency for teams deploying agents and scalable AI applications.

3. Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio

Source: Latent Space Link: https://www.latent.space/p/lab-in-the-loop-at-compute-speed

The article features Ian Quigley of Leash Bio discussing how the company tamed DELs, an assay that “traumatizes everyone,” and used the resulting data to catch cheating models. It also explores why that same data is well suited for AI agents learning to think like chemists.

This is an applied AI case study at the intersection of lab automation, data generation, and model evaluation. For builders, it shows how tightly coupled experimental loops can expose model failure modes and produce domain-specific training data for scientific agents.

Impact Analysis: Lab-in-the-loop pipelines can turn difficult experimental work into a defensible data advantage for scientific AI agents.

4. Crossing the Threshold: How OpenAI and Anthropic Approach Autonomy and Safety in Financial Agents

Source: Latent Space Link: https://www.latent.space/p/crossing-the-threshold

At AI Engineer New York, frontier labs and financial firms explored how to give agents useful autonomy while keeping consequential decisions under control, according to Latent Space. The discussion focused on how OpenAI and Anthropic approach autonomy and safety in financial agents.

Financial services is a high-stakes environment for agent deployment, so the patterns discussed around permissions, oversight, and escalation matter for broader workflow automation. Builders can adapt these controls for agents operating in regulated or high-impact domains.

Impact Analysis: Agent autonomy in finance will depend on guardrails that allow useful action while keeping humans accountable for consequential decisions.

5. The marketing automation model that drove 9x qualified lead growth in a secondary market

Source: HubSpot Link: https://blog.hubspot.com/marketing/marketing-automation-secondary-market-playbook

HubSpot reviewed its Japan marketing automation programs and found automated emails generated about 12 qualified leads per month against a goal of 160. The article describes the marketing automation model that ultimately drove 9x qualified lead growth in that secondary market.

This is a concrete applied automation case study for teams using AI, email, and funnel workflows to scale demand generation. It points to localization and market-specific execution, not just sending more automated email, as the source of improvement.

Impact Analysis: Automation gains come from redesigning the full funnel for local market conditions, not simply increasing automated outreach volume.

6. Claude skills vs. plugins vs. MCP connectors: A simple guide

Source: HubSpot Link: https://blog.hubspot.com/marketing/claude-skills-plugins-mcp

HubSpot published a guide distinguishing Claude Skills, plugins, and MCP connectors, aimed at users who open Claude’s Customize settings and are unsure which option fits a task. The article promises a simple explanation of the differences.

As AI assistants become workflow platforms, understanding these extension types is essential for automating tasks and connecting external tools. Builders can use the guide to choose the right integration pattern for content, data, and actions.

Impact Analysis: Choosing the correct Claude extension model can reduce integration friction for AI-powered workflows.

7. AI is mass decompiling games. That's both awesome and scary.

Source: Simon Willison Link: https://simonwillison.net/2026/Oct/14/ai-is-mass-decompiling-games/

Simon Willison writes that AI is mass decompiling games, a development he calls both awesome and scary. The article frames the trend as a notable AI-driven project phenomenon.

This illustrates how AI can automate reverse-engineering work that previously required scarce expert time. For builders, it raises questions about legality, intellectual property, and preservation while showing AI’s power to accelerate complex technical projects.

Impact Analysis: AI-driven reverse engineering can unlock legacy software but will force teams to navigate legal and ethical boundaries.

8. Tracking AI Overviews: What the data actually shows

Source: HubSpot Link: https://blog.hubspot.com/marketing/tracking-ai-overviews

HubSpot examines how AI Overviews have shifted audience discovery, with answers appearing directly on search results pages. The article says brands are either cited or ignored without users ever visiting a website, and it presents data on tracking these outcomes.

For AI builders and content teams, this is a case study in measuring visibility in AI-mediated search. It affects content strategy, brand monitoring, and the analytics workflows that automate reporting on AI citations.

Impact Analysis: Teams need to track AI citations as a first-class marketing metric, not just traditional search clicks.

Source Links

• Simon Willison - Anthropic tells investors quarterly revenue up nearly 50% to $16.8bn ahead of IPO
• Latent Space - Modal’s $1B ARR, $16B Series D makes it NYC’s newest AI Infra Decacorn, 3x usage and 3x valuation in 5 months, 15x in 1 yr
• Latent Space - Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio
• Latent Space - Crossing the Threshold: How OpenAI and Anthropic Approach Autonomy and Safety in Financial Agents
• HubSpot - The marketing automation model that drove 9x qualified lead growth in a secondary market
• HubSpot - Claude skills vs. plugins vs. MCP connectors: A simple guide
• Simon Willison - AI is mass decompiling games. That's both awesome and scary.
• HubSpot - Tracking AI Overviews: What the data actually shows