The idea Google is threatened by AI makes zero sense and this is why.
ChatGPT feels a lot like Alexa did.
It's astonishing, magical, profound, and awe inspiring on first, second and third use. But.....
A few days later and you're back to old habits.
Perhaps it's you and not being able to retrain your brain to do things differently, habits are hard to form, brain workflows are surprisingly non plastic. I once tried to write a chapter in my book by using Voice to Text, and I realized it's not like channeling words through your
But perhaps it's the technology.
Perhaps it can be both incredible but superficial. Perhaps the first hours you find out what it CAN do, as you explore it.
Then the following days, you find out what it CAN'T do, by using it for real world use cases. The limitations are real and rather firm. The disappointment is palpable.
I'm deeply curious by new technology, I love to play with it, to dream about it, to focus on what it means to people.
So the last few weeks I've been trying to use GPT3 wherever possible, and rather like Alexa, that means not that much.
- Do I want Google to find a Christmas Dinner recipe or do I want to create one with AI?
- Do I want Google to find the best way between São Paulo and Trancoso, perhaps with real time info, prices and availability, or do I want to generate an answer with AI?
- Do I want AI to tell me if my sore throat is bacteria based or viral?
-Is Google the best way to find the cheapest brand of OLED TV?
There are some rather key differences.
Most of the time we don't want to create a new answer to a question never asked, we want to find a brilliant existing trustworthy proven answer to a question many have previously pondered
Answer generation is an oddly niche need. One deeply valuable, but not easily monetized.
Asking Google for opening times and dates to the NRF Big Show in NYC, means advertisers can gain by delivering me contextually relevant, helpful, real time information in addition to the information I'm asking for. Google can surface me both the opening times, but also show me ads for nearby hotels, for flights to get there.
But the most useful queries on Generative AI may be to write a song about the Abilene Paradox, or to describe quantum computing, or to show me market research for Pet-food in Saudi Arabia, or develop a coffee shop business plan for a small outlet in Miami Beach. These are very valuable replies, but not for advertisers.
In short what is CURRENTLY lacking in GPT3, before huge advances are promised in later editions are.
1) Trust, we don't yet have faith in the answers like we do a source on Google.
2) Immediacy, GPT isn't connected to the internet, so answers are based on data sets a few years old.
3) Originality. AI can't think, it can only render out combinations of existing thinking based on logic and relationships between words.
Now things will change, GPT will improve and we will alter how we think to start
Comments