← All episodes
Daily AI Brief

Testing ads in ChatGPT

Wednesday, May 6, 2026 · 25 stories · 4:20 · 532 words

On Wednesday, May 6, 2026, the AI landscape saw a flurry of activity, with major players like OpenAI, Google, and NVIDIA making significant announcements. It's a mixed bag today, with updates on everything from ad testing in ChatGPT to advancements in serverless compute and AI-assisted code migration.

OpenAI has begun testing ads in ChatGPT, with a focus on clear labeling, answer independence, and strong privacy protections. The company is also introducing a new safety feature called Trusted Contact, which notifies a trusted individual if serious self-harm concerns are detected in a user's interactions with the platform.

In another development, OpenAI has partnered with Singular Bank to build an internal assistant using ChatGPT and Codex, which helps bankers save 60-90 minutes daily on meeting prep, portfolio analysis, and follow-up. The company has also announced that Uber is using OpenAI to power AI assistants and voice features that help drivers earn smarter and riders book faster.

On the infrastructure side, Databricks has published a post on rethinking distributed systems for serverless performance and reliability, highlighting the challenges of building truly serverless compute for Apache Spark. Meanwhile, NVIDIA has announced the NVIDIA Spectrum-X, an open, AI-native Ethernet fabric that sets the standard for gigascale AI, now with MRC.

In funding news, Elon Musk's Texas chip factory could hit $119 billion, according to recent filings, while QuTwo, a Finnish AI lab, has closed a €25 million angel round at a €325 million valuation. Google has also acquired a minority stake in CCP Games, the developer behind Eve Online, in a deal valued in the millions of dollars, which will allow DeepMind to train AI on the game's complex social systems.

A quieter note from dbt Labs: the dbt Developer Agent is now in preview, providing a coding agent for analytics engineering that helps users ship faster without breaking downstream. ClickHouse, on the other hand, has published a post comparing the write-side cost of Snowflake and ClickHouse under continuous ingest, showing how ClickHouse obtains query-ready data at 22× lower cost.

Meanwhile at Google, the company has published a case study on how Breuninger, a fashion and lifestyle company, boosted sales with its "be your own model" AI, which helps online fashion shoppers see how clothes would look on them. Google has also announced a new feature called The Blueprint, which highlights how customers are tackling unique challenges across industries using the latest AI and cloud technologies.

And one for the security folks: Braintrust, an AI evaluation platform, has triggered a full credential rotation after detecting unauthorized access in one of its AWS accounts, which contained customer API keys used for authentication. Apple has also agreed to a $250 million settlement in a class-action lawsuit that accused the company of misleading customers about the AI capabilities of its Siri assistant.

On the research side, Hugging Face has published a post on vLLM V0 to V1, correctness before corrections in RL, while AlphaEvolve has announced its Gemini-powered coding agent, which is scaling impact across fields. Snowflake and Veeva have partnered to unlock agentic AI in life sciences, connecting Vault data to power secure analytics and cross-functional insights.

That's the Daily AI Brief for Wednesday, May 6, 2026. More tomorrow.

Sources covered

1 story · 25 sources (6 news, 19 company, 0 voices)

  1. Testing ads in ChatGPT