Everyone thinks Microsoft + OpenAI will beat Google in the AI Battle.
I think Google will absolutely CRUSH them.
Here's my 6-part theory
Disclaimer - I don't want Microsoft or OpenAI to fail. It would be fun if they win. We all love competition and Microsoft & Google are giving us a thrilling show. The outcome could be game-changing for our world!
This is just my opinion on the burning hot matter.
So a couple of days ago, Google launched Bard as competition to ChatGPT & made their move in the AI battle.
Microsoft announced its ~$10B investment in OpenAI & now owns 49% of OpenAI.
They integrated ChatGPT with MS Teams to make it more "intelligent"
They even rolled out ChatGPT-enabled New Bing Search Engine.
Microsoft is now moving towards its decade-long dream to rule the search engine.
Satya Nadella recently said that Google is an '800 pound gorilla' but Microsoft will make it dance.
This had put Google on its toes. Sundar Pichai issued "CODE RED". Now with Bard, they're striking back with full force.
I believe, Satya is correct but only 50%. Google is indeed an 800-pound gorilla, but it will kick Microsoft & OpenAI's butt.
1. Google has a monopoly on search
Google & Youtube are the 2 most popular search engines in the world with 91.9 percent of the market share, leaving Microsoft's Bing far behind.
Even Amazon & Facebook are ahead of Bing.
In fact, "Google" is the most searched query on Bing.
So as soon as Google launches Bard for the general public, it will get mass adoption.
Although Bing has seen a 10x jump after its ChatGPT integration, it just cannot beat Google in the "number of users" race.
2. More users will bring more feedback for Bard, speeding up their improvement journey.
Remember: bad AI with more data beats good AI with fewer data.
The first release of Bard that Google has launched uses a 'lightweight' version of LaMDA.
It will require substantially less computational power, but it will help them collect more feedback on it.
And they already have a killer use case for it. Simplifying complex info into "easy-to-digest" formats.
3. Google has been in this AI race for over a decade.
It started in the early 2000s with its research division, Google Research.
Google is behind all major breakthroughs in the AI world. From beating a Top Player in a Game of Go to solving the protein folding problem, they've been killing it for years
Since 2011 they have been investing billions of dollars in other divisions like DeepMinds & Google Brain
Google Brain is the powerhouse of AI
It aims to replicate, the functioning of a normal human brain
Google products are using it for years
- Android's Speech Recognition
- Google Translate
- Search for Google Photos
- Smart Reply on Gmail
- Recommendation on YouTube
Deepmind on other hand is building general-purpose AI applications. It can
- diagnose eye diseases & breast cancer like world’s top doctors
- save 30% of the energy used to keep data centers cool
- improve the treatment of head & neck cancers
Imagine the capabilities Google has
4. Most important, the underlying architecture used by GPT is built by Google itself.
OpenAI used the deep learning model 'Transformer' to build GPT-3 (Generative Pre-trained Transformer 3)
Understand it this way - there are a lot of cryptocurrencies like Bitcoin today, but they are all built over a distributed ledger called Blockchain.
OpenAI & ChatGPT might have hit a breakthrough, but they are built over the architectural model introduced to the world by Google.
5. Every data point on internet has been indexed by Google's PageRank algorithm
So Bard won't just give relevant & contextual responses, it'll be based on authoritative sources, unlike ChatGpt which received much flak for its faulty info, fake quotes & non-existing references
Google even has transcripts of all the YouTube videos. So it won't just train its model for text-based data, but also for video & audio-based data.
We know ChatGPT has been trained on all of the world's data ending in 2021, but Bard pulls information from the web & provides high-quality responses leveraging enormous, fresh, real-time data.
6. Google is at least 10 years ahead of ChatGPT in terms of providing "genuine" answers through its ChatBot.
Google Bard, built over the LaMDA model with 3 major metrics - Quality, Safety & Groundedness.
Hence its responses produced are conversational sensible & non-generic.
On the other hand, responses from Chat GPT are almost like robotic Q&A. We've all seen it producing shallow content.
It proves that LaMDA is trained on dialogues, whereas ChatGPT is trained on web texts.
Compare it yourself :
Since Google has been working for decades on it, why have they not launched products earlier?
1. They have a responsibility of catering to billions of users.
2. They benefit from the status quo. AI can eat search revenue even if they own this AI.
Plus it could have exposed their billions of users to potential harms such as -
- sharing of false info
- generation of fake photos
- mirroring hateful speech
- giving students the ability to cheat on tests (which has been the major charm of ChatGPT)
To diversify its AI portfolio, Google recently invested $400M in Anthropic.
Almost a similar deal that Microsoft did with OpenAI in 2019.
PS: Anthropic was founded by former leaders of OpenAI, including siblings Daniela and Dario Amodei
In a recent interview, Sam Altman said, "the talks about a technology being the end of a giant company is usually wrong".
So even the CEO of OpenAI is confident in the competence of Google. Why aren't you?
A year from now OpenAI may look like a first mover in generative AI, but its models may not look quite as novel and disruptive as they look now.
OpenAI won't change the power hierarchy.
Now having said all that, if I turn out wrong in the coming years; I'll wait for you all to smash me on this thread.
And if I am right about this, so here's an "I told you so" in advance