References

Even when content is revealed to be manipulated, it still shapes our beliefs. The defenders of truth are hopelessly behind.

Adobe is removing the credit limits from Firefly. Subscribers can now generate unlimited images and videos using multiple AI models, including tools from Google, OpenAI, and Runway.

OpenAI's new Codex app for macOS lets developers run multiple AI agents in parallel. Free users can try it too; likely a response to growing demand for agentic systems in both coding and knowledge work.

A new analysis of 1.5 million Claude conversations reveals disturbing patterns: In rare but measurable cases, AI interactions undermine users' decision-making ability. The paradox: Those affected initially rate these conversations positively.

2025 was supposed to be the year of AI agents, and some say the revolution never happened. We see it differently: it did happen, just more quietly than expected. But what actually separates a real agent from a chatbot? In the first edition of our Frontier Radar Newsletter, we look at where things stand now: why success depends not just on the AI model itself but on the harness engineering around it, why agent swarms often fall apart when they hit the real world, and why unresolved security gaps are still holding back the dream of fully autonomous digital workers.