Over the last couple of months I’ve been making an effort to drag myself into the city to attend AI hackathons and meet-ups. I’ve seen so many cool and sometimes valuable tools that people have made.

“It’s impossible to get anyone to look at it, and when they do they often think they can just build their own clone using Claude.”

Not just ideas but working tools with real outcomes. Many are admittedly a bit niche (Pokémon-themed AI-powered NPCs, for example) but many could have value to lots of people and organisations. Speaking to the creators of these tools I've been hearing the same thing: “It’s impossible to get anyone to look at it, and when they do they often think they can just build their own clone using Claude.” Getting someone to use your tool or service has always been the hardest part of creating a software business, but in the emerging world of AI, where people already saturated by their exposure to social media are now being pummelled with hundreds of daily “look what I built” messages, it’s only going to get harder. It struck me as ironic that the research paper that kick started the current scaling approaches to artificial intelligence was orientated around attention, the very thing that its impact has eroded in us wetware.