With all the hype around OpenAI and ChatGPT, I thought it might help to curate a little no jargon summary of the main ‘wins’ that ChatGPT has created and also transparently highlight where the gaps / risks still are and what the future may hold.?References are all listed at the bottom.?
- Generative AI (GAI) has been around for a long time.?ChatGPT is the latest in a long list of other GAI apps.?But it is the consumerization enabled by the unprecedented conversational capabilities and ease of use that led to the recent hype (remember Lensa?)
- Microsoft recently invested $10 billion in ChatGPT’s developer, OpenAI
- Google is releasing its own AI tool, “Bard”
- Chinese tech giant Baidu is preparing to launch a ChatGPT competitor
- Other alternatives to ChatGPT:?Jasper Chat, Bing AI, Chai, Chatsonic, Chinchilla (by Deepmind)
- There is a huge landscape of Generative AI apps out there (currently 663 other apps) in various fields
Chat GPT + Knowledge mining
- ChatGPT is a natural language processing tool that can answer specific natural language questions (in over 100 languages), and assist users with tasks such as composing emails, essays, and code (including Python, Javascript, and C++).
- Despite the hype around ChatGPT — and Generative AI (“GAI”) overall — there are major practical, technical, and legal challenges to overcome before these tools can reach the scale, robustness, and reliability of an established search engine such as Google
- Google Search has been successful in finding websites that have the information you are looking for (“Search 2.0”).?But there seems to be a new use case on the rise now. Users want “deeper insights and understanding.” (i.e. “Search 3.0”)
- Search 3.0 delivers answers instead of websites.?If Google has been the colleague who points us to a book in a library (‘the web’) that can answer our question, then ChatGPT is the colleague who has already read every book in the library* and can answer our question. But, there are major challenges …
*Key challenges with GAI, LLMs and Chat GPT:
- In its current form, ChatGPT is not a search engine, primarily because it doesn’t have access to real-time information the way a web-crawling search engine does. ChatGPT was trained on a massive dataset with an October 2021 cut-off. As far as ChatGPT is concerned, Russia hasn’t invaded Ukraine.
- Continuously retraining an LLM as the information on the internet evolves is extremely difficult.?The most obvious challenge is the tremendous amount of processing power needed to continuously train an LLM, and the financial cost associated with these resources.
- LLMs are trained on carefully selected datasets that the developer deems to be appropriate. But this level of curation does not ensure that all the content in such massive online datasets is factually correct and free of bias, nor is it representative of all data that exists.
- Where in "Search 2.0" users apply their judgment to distinguish fact from fiction, opinion from objective truth, and decide what information they want to use, the judgment-based step is removed with ChatGPT, which makes it directly responsible for the biased and racist results it may deliver.?Users have no idea what sources are behind an answer with a tool like ChatGPT.
- Generative AI has the potential to transform entire industries, but its cost and the required expertise puts the technology out of reach for all but a select few companies (CIO dive).
- LLMs will likely enhance certain aspects of traditional search engines, but they don’t currently seem capable of dethroning Google search, (nor critical thinking analysts or management consultants!).
- They could play a more disruptive and revolutionary role in changing other kinds of search, knowledge mining and the ways in which we all source information in the future – especially if trained on wider sources.?
- What is more likely in the “Search 3.0” era is the rise of purposefully and transparently curated and deliberately trained LLMs for vertical search, which are specialized, subject-specific search engines.
- There will be new “ChatGPTs” and “Lensas” popping up all the time, let’s see how it develops and which ones really scale! (There will also undoubtedly be a wave of announcements of partnership between consulting firms and AI companies, but I suspect that the ones that aren't "announcing" are already long in partnership with the big names.
- Some of the winners of the GAI revolution will be chip manufacturers and infrastructure providers to train GAI networks – as GAI requires massive compute power to train.
very helpful curation
Knowledge Management Leader / Community Manager / Change Agent / Mentor & Coach
2 年Thanks, Chris. Appreciate the no jargon! Have shared with some teammates.