Nicholas Thompson’s Post [Video]

Hi Nicholas Thompson, at Neeva we are working on search that combines the goodness of AI models with the authority and timeliness of search. Check out what our AI generated for your query. More on this coming soon!

I agree. Perhaps now, bot AI responses are balanced and well supported. but I don't think they will stay this way when answers become business. When this moment arrives I won't trust a bot response the same way I don't trust a post with no references. Human-made verifiable helpful content is irreplaceable and seach engines challenge is to recognize and rank it.

Many of the answers provided by ChatGPT are wrong and dangerous.

Clear readable responses that are inaccurate or deliberately misleading render the entire endeavour worse than pointless. Worse than pointless because they will undermine trust in reliable and reputable sources. This is a huge threat to democracy and to society.

If chatgpt and generative ai applications based on LLM are to responsibly scale, they HAVE to figure out standards/best practices for eliciting diverse user feedback for content moderation purposes (will require different approaches than legacy Reddit or Wikipedia because crowdsource is now solo source in a 1:1 conversational interface). A closed loop system where there’s output from black box ML training materials poses huge risks (bias, disinformation, security vulnerabilities, etc) .

Do you want to know WHY #ChatGPT is not suited for business-grade conversations? Why Generative AI ≠ Conversational AI Read on. https://www.linkedin.com/posts/srinipagidyala_chatgpt-not-for-business-activity-7005331631754399744-heit?utm_source=share&utm_medium=member_desktop

Here's what is said about Agile Software Development #agile Agile software development is a type of software development methodology that focuses on the continuous iteration of development and testing throughout the software development process. The goal of agile software development is to deliver high-quality software quickly and efficiently by bringing together small, cross-functional teams to work on multiple aspects of a project simultaneously. This approach to software development is based on the principles outlined in the Agile Manifesto, which emphasizes collaboration, flexibility, and transparency. Agile software development is often used in the context of large, complex projects that require frequent communication and coordination among team members. It is designed to help teams respond to changing requirements and customer needs, and to deliver working software quickly and iteratively.

I’m pretty confident Google will experiment with including this type of tech into its ”one box” results, i.e where they already try to provide an answer (with links/citations) directly in the search results. But the source issue is a big one, not least because the whole Google ranking system is based on the concept of trust/authority. Without the souce I guess it would also be a possible (legal) minefield in terms of the advice given; would this mean Google is now a publisher and not just a gateway to relevant content?

So are we saying that if the model uses information from doubtful sources, hoaxes, or skewed information, we get skewed or even entirely inaccurate results? Does it mean that such chatbots could actually fuel a new wave of hoaxes and "propaganda," which will be even tricker to expose as they might be more elaborate or comprehensive? Kind of surprising :)

Lol the source list of your average ChatGPT answer will make Microsoft’s terms of service agreement seem like a haiku.