🚨NEW: The confirmed list of social media apps/sites under-16s in the UK will be banned from using: - TikTok - YouTube - Snapchat - Instagram - X (formerly twitter) - Reddit - Facebook - Twitch - Kick - Threads

@KanishkaNarayan has a Jekyll and Hyde personality: By night, he is seducing startups with a vision of a great AI-first UK. By day he's on television sprouting unspeakable nonsense about saving children by installing spyware on millions of phones, about needing to do ID checks on UK citizens, about taking down US companies. So here's the thing: Either he has no clue technically - which is very worrying given his AI job - or he knows it's all nonsense, he knows it's child safety masking a state surveillance plan. And he know it's technically impossible to work, it's a security disaster, and it's going to lead to a good fraction of UK citizens intentionally working around the law because they don't want to be mass surveilled. And, given the experience in Australia, he knows it's going to be completely ineffective. I guess it takes a politician to hold two such irreconcilable views and try to push them both. It's hard to give credit for the Dr Jekyll side of the role, given the intentional misinformation being sprouted by Mr Hyde.

Reform is importing populist tactics which have undermined democracy elsewhere to here, fuelling online mis and disinformation, conspiracies and murky crypto donations. We need tougher action. My latest in the Guardian 👇 theguardian.com/politics/202…

Completely agree. The way you turn AI into a compounding asset is by codifying your best ways of working and then democratizing it across your company in a way that wasn't possible before. One of the more common requests we get from companies earlier in their AI journey now is to stand up their internal skill library & fill that library with skills that we build through interviews/sessions with top performers.

Keir Mather showing why he’s the youngest minister since Gladstone

Superb intervention from Tony Blair. Agreed with almost every aspect of his analysis.

Just added Skills to performance.dev/skills! Every article now has a Markdown version built for agents. Over time, I’ll be curating a living performance skill from everything I’ve learned along the way.

Consulting firms have been lying to themselves and their customers for months. They keep charging for the pre-AI length of missions while their consultants are usually able to complete them in 1/3 of the time using AI. They have historically been selling top 1% human intelligence at a very high price but AI is already able to do most of the work their consultants do today, for a fraction of the cost and time. Their customers know that hence their doubts on the current value proposition. They will not admit it, but human consultants are the bottleneck. Most firms are trying to augment consultants instead of redesigning the work. In discovery, for example, consultants still run the interviews manually, record them, and feed the transcripts into an AI model to structure the insights and draft recommendations. That makes each consultant faster but does not change the scale of the mission. They can still only interview as many people as human time allows. Today’s models can already do a large share of the work consulting firms charge millions of dollars for. And that share will only grow. There is no reason for humans to do work that agents can already do faster, better, and at scale. One of the usual objection is domain expertise that humans and consulting firms have built over time, but that is mostly a temporary and feedback loop problem. Humans should step in at the end of the loop, let the agents do most of the work, review their outputs, correct reasoning, and feed their expertise back after every mission in order to then build deeper industry and workflow-level knowledge over time. Mission after mission, this will become a dataset no human team will ever match. We are rebuilding the AI-native version of consulting at Foaster and our customers who used to hire MBBs love it.

Introducing performance.dev! A new space where I explore how the best apps in the world are built. First piece: How's Linear is so fast? a technical breakdown. performance.dev/how-is-linea…