THE RUNDOWN

On a lighter note, OpenAI is rolling out a shopping feature that lets ChatGPT users upload their own photos and see how clothing and accessories would look on them. You can also save items you like for later. It's a direct play for e-commerce attention, and it puts ChatGPT squarely in competition with fashion apps and even Google Shopping.

A federal judge has dismissed antitrust suits from Chegg and Penske Media Corporation, both of which argued Google's AI Overviews were killing their web traffic by answering questions without sending users to their sites. The judge agreed that AI search has real consequences for publishers, but ruled it's not an antitrust violation. It's a significant win for Google — and a sobering signal for media companies banking on search referrals.

Three OpenAI employees were let go after allegedly sharing information with an external AI safety organisation. The irony isn't lost on anyone: a company built around safe AI development firing people for apparently caring about AI safety. Details are still thin, but it's raised eyebrows across the industry about how openly OpenAI actually engages with outside oversight.

US federal cybersecurity has had a rough October already. Two government agencies were breached within weeks of each other, spilling a significant amount of sensitive data. The back-to-back nature of the attacks suggests either a coordinated campaign or serious systemic weaknesses — and it's a reminder that even the most resourced governments struggle with basic digital defence.

The founder of Mandiant — one of the most respected names in cybersecurity — has a new venture called Armadin, and it's already valued at $2.5 billion after raising $255.5 million. The approach uses swarms of AI agents to probe and protect enterprise systems. Given Mandia's track record, this one is worth watching closely.

California has passed a law that will fine autonomous vehicle operators when their cars block first responders. It's a direct response to real incidents where driverless taxis have sat frozen in the middle of roads during emergencies. As robotaxi fleets grow, regulators are clearly done waiting for the industry to self-correct.

Sony is launching Quick Spectral Super Resolution (QSSR), an AI-powered graphics upscaling technology for the regular PS5 — not just the Pro. It promises better image quality without the hardware upgrade cost. For gamers in Ghana and across Africa who haven't upgraded to a PS5 Pro, this is genuinely good news.

Apple is reportedly working on a home security camera that skips video recording entirely and instead sends text descriptions of events — think 'person arrived at door at 6:14pm' rather than actual footage. It's a privacy-first design that's very Apple, though whether it'll satisfy people who actually want to review footage is another question.

Lyft has agreed to settle a lawsuit from 2020 over how it classified its drivers as independent contractors rather than employees. The $272.5 million settlement closes a years-long legal headache, though it doesn't change the current status of gig workers. The broader battle over contractor classification in the gig economy is far from over.

Android Central, one of the longest-running Android news blogs, laid off its entire editorial team — but owner Future says the site will keep publishing. That's either very good news or a sign the site will pivot to AI-generated content. Either way, it's another rough moment for tech journalism, and worth paying attention to as a trend.

That's your October 2nd briefing. From AI chatbots influencing foreign policy to silent security cameras and a fashion-forward ChatGPT — tech is moving fast, and not always in the direction anyone planned.