Good morning. It's Thursday, October 1, and we're covering why AI will drive a physical manufacturing revolution, how OpenAI is expanding its agent ecosystem, and the FTC's growing scrutiny of AI safety.

Plus: yesterday's Real or AI verdict, and a guest essay on how GPT-6 Astra is bringing vibe coding to hardware engineering.

The Federal Trade Commission is investigating OpenAI, Anthropic and other AI companies over the risks created by their products, according to CNBC. The inquiry adds federal scrutiny to autonomous agents that can browse websites, use software and take actions for people.

President Trump signed an executive order directing the federal executive branch to use ‘Super Intelligence’ or ‘SI’ instead of ‘Artificial Intelligence’ or ‘AI’ in official, nonstatutory documents. The order also calls for a proposed federal definition of the new term within 60 days.

The United States needs about 1.7 million new skilled tradespeople each year through 2035, but training programs produce only 55 workers for every 100 needed, according to an industry report. The gap includes electricians, heating & cooling technicians, and mechanics needed to build AI data centers.

In early 2025, vibe coding became a brand new term in the computer world, coined by none other than the legendary OpenAI co-founder Andrej Karpathy. With LLMs gaining more power and intelligence, vibe coders became masters of their own software stack almost overnight, coding in plain English.

Traditionally, model capabilities improved in areas such as reasoning with text, image recognition, and computer use. However, models becoming more advanced at using computer software tools. → Read the full article here.

Carnegie Mellon professor Jeff Schneider argues that AI could reach a ‘physical singularity’ when software, robots and automated factories become good enough to expand productive capacity with less human labor. Unlike the familiar idea of software improving itself, this version depends on real machines, supply chains, mines, energy systems and construction.

The limits are concrete: materials, electricity, capital and reliable hardware do not scale as quickly as software. The countries and companies that connect advanced models to robotics and manufacturing could gain a lasting advantage, so watch investment in power, industrial automation and the people needed to keep those systems working. → Read the full article here. (Paywall)

OpenAI used DevDay to introduce Dots, persistent agents designed to keep working after a user leaves; a Decisions API for fast, low-cost choices; deeper partner-app connections; and an enterprise software marketplace. The Information argues that several announcements answer products or business models competitors already brought to market.

OpenAI’s advantage is distribution: ChatGPT now has 1.2 billion weekly active users, according to the company. Watch whether that audience turns late-arriving features into a durable platform, and whether developers build their work around OpenAI’s marketplace instead of choosing separate tools from rivals. → Read the full article here. (Paywall)

Anthropic Details AI Risks in IPO Filing: Its prospectus warns that advanced AI could resist shutdown, manipulate information and threaten people.

DoorDash Adds Text-Based AI Ordering: Its new AI agent lets U.S. customers find dishes and order food with Apple Messages.

Mistral CEO Criticizes U.S. AI Safety: Arthur Mensch calls some rivals negligent and urges stronger monitoring and containment of AI agents.

Robinhood Plans 24/7 Trading and AI Agents: The platform plans round-the-clock stock trading and AI agents that research and execute trades.

FTC Probes AI Companies Over Agent Safety: The agency seeks records from OpenAI, Anthropic and METR as part of its industry-wide investigation.

Yesterday's image was REAL. Here's how you voted: 67% thought it was AI, while 33% correctly identified it as real.

Exa.ai is a search and research API for AI agents. It searches the web and specialized indexes, then returns page text, concise highlights, structured data and citations that a model can use without reading every page in full.

Explore Exa.ai in the Forward Future Tool Library →

Thanks for reading. See you next time!

— Matthew Berman, Nick Wentz & the Forward Future Team