

Yeah, the article cites that as a control, but it’s not at all surprising since “humanity by survey consensus” is accurate to how LLM weighting trained on random human outputs works.
It’s impressive up to a point, but you wouldn’t exactly want your answers to complex math operations or other specialized areas to track layperson human survey responses.
This is an incredibly undeservedly positive profile for this company. So many paragraphs here look like they were written by Bending Spoons or are uncritically repeating direct claims by the company.
“Transform them to serve millions of users more efficiently”? “Its main focus is on making improvements to products and services…”?
Awfully strange ways to say their business model is “find companies with entrenched or locked-in userbases and then turn goodwill and inertia into profit by laying off all but a skeleton crew, increasing prices and lowering services and quality.” Good thing that is totally different from what a private equity buyout does.