ChatGPT has ignited the frenzy about the future of AI. Not only have consumers flocked to try the generative AI service, but also have entrepreneurs and investors trying to latch on the service in hopes to to build the next billion dollar companies.
According to our friends at Measurable.ai, the number of ChatGPT users has already surpassed 120 million, with more than 3 million using the service daily.
Excited, people are also imagining how ChatGPT, or generative AI in general, is changing the dynamics of big techs.
Microsoft is clearly a beneficiary in this, having backed OpenAI, the company behind ChatGPT, for $1b as early as in 2019. Just this week, it announced a series of new initiatives integrating ChatGPT-based services into its Edge browser and Bing search engine. It has also recently struck a quite smart deal to invest an additional $10b into OpenAI.
Not only FOMO, but existential threat
The other giants are feeling the pressure if they miss out the opportunities of generative AI. We know that a few have developed models similar to GPT3, the one that powers ChatGPT. Notably Google has been making a lot of investment into AI because the company’s cash cow – search and advertising business – depends on it.
It was reported that Google was in panic mode (“code red”) when ChatGPT was launched. Google probably rushed the launch of Bard this week – their answer to ChatGPT based on their own LaMDA model.
However, as you can clearly see, the market did not respond very positively to Bard:
Knowing Google, we believe they will see ChatGPT as an existential threat and will throw all the companies’ resources on improving Bard to be at least on par. The question, obviously, is are they too late? Do they still have a chance?
If you ask ChatGPT this question, their answer goes like this:
As you can see from the above, an amazing feat of ChatGPT is its ability to steer itself away from polemic and toxicity. The response – as well as many we have tried – sounds like what a seasoned human professional would have said.
After being trialed by more than 100m users, ChatGPT has generated a lot of funny responses people share on social media, but it has not said something too edgy that would backfire. Getting to that point where OpenAI felt it was comfortable to release ChatGPT to the open probably took years of training.
Has Google done that, even if LaMDA is as good as if not better than GPT3? Will Bard withstand the same scrutiny of billions of uncontrolled questions from hundreds of millions of users? Will users actually bother with Bard if ChatGPT is already good enough?
These are all the questions that would probably keep Google awake at night. They do not have a choice – though – but have to carry on making improvements to Bard (or any other generative AI-based service).
The parallels with Tesla
In a way, I almost feel that there is some parallel between OpenAI and Tesla – as far as Google is concerned.
Remember in the early days of the craze about autonomous driving – Google and Tesla took very different approaches:
Google created its own unit (later spun off to become Waymo) to develop L4 (advanced level of autonomous vehicles defined by SAE International’s J3016 standard). It takes years to make the system good enough, and one accident to set back the whole project for months if not more years.
Tesla, on the other hand, went early with its L2 autopilot already embedded in millions of cars it sold. The system gives Tesla real-time feedback from millions of users – and improves its autonomous driving capabilities.
Who will reach real L4 commercial success? It seems like Tesla’s approach is, for now, winning.
Now ChatGPT has that advantage over Google Bard. Not only that, its system is better as well.
Of course, as ChatGPT will tell you – it is very hard to predict the future. But there are some fundamental lessons here we can learn from the evolution of ChaptGPT versus the undertakings of other big techs.