The moment the news about Microsoft'™s investment in OpenAI came out, the debate, ChatGPT over Google Search, broke the internet.
So far, most of our conversations have been incomplete, missing key points on both sides.
Therefore, we decided to look at the events from a macro and micro perspective.
So here is what happened.
In the third phase of their long-term partnership, Microsoft announced a multiyear, multibillion-dollar investment in Open AI to accelerate AI breakthroughs and ensure that these benefits are broadly shared with the world.
This follows their previous investments in 2019 and 2021. The partnership will bring several exciting developments including:
Supercomputing at scale - Microsoft will increase its investments in the development and deployment of specialized supercomputing systems to accelerate OpenAI’s independent AI research. Additionally, Microsoft will build out Azure’s leading AI infrastructure to help customers build and deploy their AI applications on a global scale.
New AI-powered experiences - Microsoft will deploy OpenAI’s models across its consumer and enterprise products, introducing new categories of digital experiences built on OpenAI’s technology. This includes Microsoft's Azure OpenAI Service, which empowers developers to build cutting-edge AI applications through direct access to OpenAI models backed by Azure’s trusted, enterprise-grade capabilities and AI-optimized infrastructure and tools.
Exclusive cloud provider - As OpenAI’s exclusive cloud provider, Azure will power all OpenAI workloads across research, products, and API services.
What’s most exciting is to see how Microsoft can infuse OpenAI's innovations to enhance user experiences in its consumer and enterprise products:
- Microsoft Word could have built-in text generation to help you write things and brainstorm essays.
- Microsoft Outlook could help you write emails and find excuses to get out of meetings.
- Microsoft PowerPoint could, of course, have AI-generated clip art and design suggestions.
- Microsoft Bing search results could be enhanced with ChatGPT integration
- Microsoft's Word, PowerPoint, and Outlook apps could be integrated with some language AI technology.
Since 2016, Microsoft has been committed to building Azure into an AI supercomputer for the world, serving as the foundation of its vision to democratize AI as a platform.
Through this partnership, Microsoft and OpenAI pushed the frontier of cloud supercomputing technology, announcing their first top-5 supercomputer in 2020, and subsequently constructing multiple AI supercomputing systems at massive scale.
These innovations have the potential to create a transformative impact, similar to that of the personal computer, the internet, mobile devices, and the cloud.
And, underlining all of their efforts is Microsoft and OpenAI's shared commitment to building AI systems and products that are trustworthy and safe.
Now let’s take a close look at the debate over whether ChatGPT will replace Google Search.
We will delve deeper into the discussion and examine the pros and cons of both Google Search and ChatGPT.
🔥 What's wrong with Search?
Search is a complex and powerful technology that has been dominant for the past 25 years, but despite its many strengths, it does have its flaws. For example, search results for "tail" queries - those that require additional information from sites like Reddit or are cluttered with SEO spam - are often less than satisfactory.
💪What can ChatGPT do better?
ChatGPT has the potential to offer a much better experience for these types of "tail" queries. It is better equipped to handle natural language queries and can offer users quick, relevant answers to their questions.
🙄 Is it too expensive?
One of the biggest concerns about ChatGPT is its cost. Estimates suggest that it could cost up to $3 per query, quickly adding up to a significant amount of money at scale. However, it is important to note that these costs are likely to decrease over time and that ChatGPT may not be the best solution for all query types.
🤔 Does it ruin ad revenue?
Another concern is that ChatGPT may harm ad revenue, as it doesn't offer the same opportunities for targeted advertising as Google Search. However, there is evidence to suggest that the opposite may be true - that by offering users quick answers to their questions, ChatGPT could increase engagement and offset any loss in ad revenue.
🤨 Is Google doomed?
Despite the buzz surrounding ChatGPT, it's important to remember that Google is still a formidable technology company with many powerful language models of its own. Google has PaLM, Flan-T5, Gopher, and other LLMs that are equally good, if not better than GPT-3, when it comes to NLP benchmarks.
🤔 Should we use LLMs as Search oracles?
Using LLMs directly as a search engine raises questions about the veracity of the information being presented, as it's impossible to verify the sources and determine the accuracy of the answers. Additionally, LLMs are non-deterministic, and Google does not want to be held responsible for spreading false information.
🤔 What can be done better?
LLMs have the potential to power featured snippets or to summarize search results pages, but the ranking of results is so effective because of machine learning models that are fed by click feedback data. Overriding these models would likely have a negative impact on the quality of the results.
🤔 Other reasons why it won't work?
In addition to cost, there are also concerns about the speed of LLMs in a production environment. Response times can vary from 2 to 15 seconds, which is not acceptable for many users. Additionally, LLMs may be more suited to use in voice assistants rather than as a standalone search engine.
⏳ Search is still relevant
Finally, it's important to remember that despite the hype surrounding ChatGPT, search is still a hugely popular and effective tool for many users. Most of Google's daily users have not even tried ChatGPT, and the "search is dead" sentiment is largely limited to a small group of highly-engaged internet users.
So, the bottom line is, while ChatGPT offers exciting possibilities for the future of search, the platform is not without its limitations and challenges.
The true future of search will likely involve a blend of both traditional search and language models, offering users the best of both worlds 🙂
