🌎 In this week’s episode of As the AI World Turns, we find OpenAI jumping in front of Google with the announcement of a bevy of dazzling new AI capabilities – but at what cost?
With Google I/O set for Tuesday, OpenAI scheduled an announcement and immediate release of its lightning fast multimodal 4o model on Monday.
The model has a friendlier – what some have called “flirty”– tone that harkens to the AI named Samantha that Joaquin Phoenix’s character fell in love with in the movie “Her.”
It sparked a renewed debate about the implications of anthropomorphizing our AI assistants.
–
Should OpenAI’s front running also spur Google to take a good look to see who in its ranks might be a spy for OpenAI?
Not that OpenAI could have instantly whipped up updates that were very similar to some of those that Google announced at its annual developer conference, but it certainly could have been a nudge for OpenAI to get their 4o model out ahead of it, and perhaps faster than planned.
If true, that speed-up may have been the last straw for two key members of OpenAI’s superalignment team, which works on maintaining the safety of its models - Ilya Sutskever (its leader) and Jan Leike.
On Friday, Jan posted on X: “I have been disagreeing with OpenAI leadership about the company’s core priorities for quite some time, until we finally reached a breaking point.”
And Jan and Ilya aren’t the only employees on the superalignment who left - at least 5 others have reportedly departed in the last few months.
–
OpenAI itself seems to have admitted they released 4o without fulling knowing its potential impact.
In its announcement around the release (link below), while the company wrote a lot about the safety measures taken, it also said:
“Because GPT-4o is our first model combining all of these modalities, we are still just scratching the surface of exploring what the model can do and its limitations.”
I guess we’ll all find out soon enough.
Kommentare