Nearly every day, there is a new update or breakthrough from a major AI company. It can be hard to keep up.
So on our Launch Codes Live show, Joe (CEO) and Lauren (VP of Marketing) took to LinkedIn Live to discuss the most important news in the world of AI so you can stay up-to-date.
Here is a brief overview of the stories they covered.
Gemini Advanced is now available for a 2-month free trial.
Google Bard is officially dead.
As of February 8th, Google’s GPT competitor is now called “Gemini Advanced” and it’s available for a 2-month free trial. After the trial is over, users will have to pay a 20 USD per month subscription to continue using it.
It’s still early, and we’re still dipping our toes in with Gemini Advanced, but the 2-month free trial is a welcomed offer that we didn’t see with Chat GPT Plus. Even as we write this, Google just made an announcement about their next generation model: Gemini 1.5.
On the full episode of Launch Codes Live here, Joe and Lauren dive deeper into how Gemini and the Google ecosystem, how it might compare to Microsoft co-pilot and GPT-4, as well as some background on how Google came up with the “Gemini” name.
Why is Gemini still only as good as GPT-4? (4 Theories)
Joe came across an interesting post by Ethan Mollick (Wharton professor and AI expert) that includes 4 potential theories on why the performance of Gemini is roughly on par with GPT-4.
Here are those theories (paraphrasing Ethan’s points):
- Gemini and GPT-4 are both at the height of what’s possible. Essentially, this is as good as it gets for AI using LLM technology.
- Google trained Gemini to meet GPT-4 levels and then stopped. More will come.
- OpenAI has some secret sauce that the other companies can’t figure out. This is just the best that Google can do.
- Gemini and GPT-4 being so similar is merely a coincidence.
Ethan thought that #2 was the most likely scenario here, but admits that he has no idea for sure.
It turns out that Ethan was right with Google’s announcement today about Gemini 1.5.
It’ll be exciting to see what comes after Gemini 1.5 and beyond.
Nvidia’s lets you run an AI chatbot on your GPU.
Nvidia recently announced “Chat with RTX”, which lets anyone run an AI chatbot locally on their GPU.
It’s still in its early stages and does require quite a significant supply of text before it becomes useful, but the ability to run an LLM without the use of the internet is another glimpse into the future of what’s possible. Eventually, we will likely see LLM’s running locally on our phones in the palm of our hand.
GPT is “less lazy” now?
Another story that came out earlier this year was the announcement that GPT should be “less lazy” now. Sam Altman said that this change is coming after an OpenAI update.
Some of the theories behind why GPT was slowing down – and even refusing to complete tasks – are pretty funny. One of them was that GPT is experiencing seasonal depression and has observed that humans “slow down” during the month of December.
The real reason for the slowdown is unclear, but it’s definitely back to normal now.
Watch the full episode.
If you missed our debut of Launch Codes Live, you can watch the full recording on LinkedIn here.
And we’ll be doing these live sessions on a monthly basis going forward, so look out for the next event on LinkedIn in March!
