Galactica is an artificial intelligence (AI) assistant that was launched by Meta at the end of last year. But why are people still talking about it alongside ChatGPT and Bard?
At the end of last year, just two weeks before ChatGPT was launched, Meta (the overlords of Facebook, WhatsApp and Instagram) decided to shoot their shot and bring an AI assistant to the internet.
Named Galactica, it was pitched as being mainly useful for scientific research. The promise was that it could write articles, solve complex maths problems, and like ChatGPT, could generate computer code, among other boffiny pursuits.
ChatGPT vs Galactica
ChatGPT is not perfect. It has been accused of political bias, and “hallucinating” facts, among other failures. However, for things where facts don’t matter — like writing a marketing plan for your online shoe collection, people have found it to be incredibly helpful. So ask it to do basic tasks, where truth doesn’t matter, and it seems to fit the bill.
Where Meta seems to have gone wrong is to launch an AI that focuses solely on Scientific research. This is where the truth really, really, REALLY matters. And it seems this might have been Galactica’s undoing.
An early adopter, for example, found that it could come up with a very authoritative-sounding article about bears that lived in space.
After three days online, Galactica was pulled down.
Many fans of the Galactica AI assistant claimed that Meta faced different challenges compared to OpenAI, the parent company of ChatGPT.
There is now a lot of talk about why it failed.
We all know that Meta will be back with a better version, so let’s hope this time it is vastly improved, or that it sticks to marketing plans for shoes.
Meta trailing behind
The Galactica failure follows on the heels of Meta’s metaverse failure. Their attempt to make their own cryptocurrency, Diem, also failed.
There were also other failures. Their facial recognition technology that automatically tagged people in pictures on Facebook was dropped after users were horrified by the invasion of privacy. (It didn’t even work very well. My friend took a photo of 12 hanging hams in Spain, and one of the hams was auto-tagged as me.)
Facebook and Instagram are losing ground against TikTok, and now, there is a new competitor in the AI space which will be a formidable opponent if Galactica is relaunched.
Google has launched their AI assistant called Bard in panic, since realising just how much traction ChatGPT already has in the market. Nobody knows if Bard will make it, but they seem to have a better chance than Meta’s offering.
Zuckerberg sticking with AI
It seems, however, that head of Meta, Mark Zuckerberg, still wants to go down the AI road despite being on the back foot. He said to investors that he would make AI development his focus for 2023. Meanwhile, the price of Meta shares plunge down a dark cavern.
For now, Galactica is offline. Despite spending billions of dollars and employing hundreds of top AI researchers, Meta has missed the boat on the current excitement around generative AI, and probably won’t get it back.
Unless they get a big win and soon, they seem to be headed into a death spiral, in the style of MySpace, which died as Facebook grew massively and pushed it out.
The circle of life in big tech sure is interesting these days.
