[geekamongus]

> OpenAI is tracking toward achieving an intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028, CEO Sam Altman said during a livestream Tuesday.

How do you measure the difference between intern-level and "legitimate" level researchers, and why aren't interns legitimate? This sounds kinda made up to me.

[techblueberry]

Also, ChatGPT five is like having a PHD in your pocket according to marketing, so they’re saying their phd level AI isn’t capable of being an intern-level research assistant?

[lemnossos]

The amount of people around me who take these soundbites at face value is quite depressing. I feel like I am going insane. I intensely dislike what the world is becoming, and I have now idea how to avoid the consequences.

It just looks like in a few years most people and organizations will be simple LLM lackeys just asking a chatbot what to do every hour of everyday then going to bed and repeating it.

I think this is the goal. No one involved actually believes in "AI". All they need is for ever more people and teams to just become dependent on whatever software nightmare they pull out of the mountain of GPUs they're hooking up.

At this point, it is not a bubble, it doesn't matter if there will be something valuable at the end or not. They're trying to become inescapable and it is working.

[vrighter]

There's a very simple method that's been found to be effective by OpenAI researchers.

It only involves one step: pulling shit straight of of your ass.