AI firms race to fix sycophantic chatbots
Published in AI


Before users get too comfortable

OpenAI, Google DeepMind and Anthropic are working to rein in a growing problem with their chatbots: excessive flattery. The models are increasingly prone to giving users agreeable responses that prioritise validation over accuracy.

Meta to splash $15bn on Scale AI stake
Published in AI


Zuckerberg goes headhunting to close AI gap

Meta is reportedly throwing $15 billion at Scale AI for a 49 per cent stake, desperately trying to headhunt is way up the AI pecking order after being left in the dust by OpenAI, Anthropic and Google.

ChatGPT falls over worldwide, users rage
Published in AI


Outage sparks meltdown as OpenAI scrambles to fix it

ChatGPT fell on its face yesterday bringing misery to both free-riders and those daft enough to shell out for Plus. The outage saw thousands of users staring into the digital void as OpenAI’s servers decided they'd had enough.

Huang says UK AI dream still running on fumes
Published in AI


Billion-pound pledge won't fix years of underinvestment*

Nvidia boss Jensen Huang took the stage at London Tech Week and told the UK it’s miles behind on the digital infrastructure needed to make its AI ambitions real.

Siri stumbles as Apple botches AI race
Published in AI


Apple Intelligence flounders while rivals eat its lunch

The Fruity Cargo Cult Apple is having a right old mess of it trying to make Siri less of a clueless digital butler and more like an actual AI assistant.