If AI can do everything, what are we supposed to do? | H-FARM

Diego Pizzocaro is a lecturer in AI topics at H-FARM College, PhD in Multi Agent Systems, CMO at the tech scaleup ForceManager, and former startupper in one of our acceleration programs.  Here are his insights – that he also shared at our CLUB Artificial Intelligence on December 2023.

In November 2023, I was honored to be invited at OpenAI DevDay, the first international developer conference organized by OpenAI in San Francisco. At the cusp of something big The event was different than any other. It felt like being in a room where every conversation potentially was shaping the future of the AI industry, which today largely represents the entire tech industry. It felt like something akin to the tech world's moon landing. AI is transforming each one of us into a developer 9 months after release, still GPT4 embodies the state of the art in AI LLMs, even after the release of Google's Gemini. The real highlight of the event was understanding that now over 100 million weekly active users can create custom chatbots (called GPTs) using chat itself. Just imagine: you have a 300 page document/book/list of LinkedIn posts/whatever, now you can pass it to a Custom GPT that you create and just tell it to use as its inspiration database for giving you the answer you're looking for. Best of all? You can share it with your friends, colleagues, or the world. The true advancement here lies in equipping a vast user base, regardless of their technical background, with the tools to develop AI agents who can do whatever they need. This is the beginning of a new era of user-generated AI applications – a shift similar to what User Generated Content has been for the Internet. The end of the "old" developers?  Absolutely no. This is a renaissance moment for the entire tech industry, especially for developers. Custom GPTs can be created at scale by developers via the newly announced Assistant APIs and can also access GPT4 Turbo, GPT-V(ision), DALL-E 3, Text-to-Speech and Speech-to-Text capabilities via APIs. Now every app can potentially become smart in no time and developers will be able to include such intelligence in every app ultra quickly. Of course they will also be able to use themselves the coding capabilities of GPT4 with the Advanced Data Analytics module or Github Copilot, or any other AI trained to code to go even faster in developing such capabilities inside their apps. Microsoft already released their platform similar capabilities and Google, Amazon, and other tech giants will follow rapidly. This is the AI race speaking loudly to developers. This advancements lead us towards a future where AI can operate with increased autonomy and precision, significantly enhancing productivity and innovation. An important note: at this stage, we cant' give AI too much autonomy because it still makes a lot of mistakes. It's still necessary to start discussing about "super intelligence", even if its arrival might still be a bit far (not sure how far). Generative AI in education and business Generative AI is already reshaping completely how we work and learn. This isn't just about following a trend. It's about harnessing real, tangible benefits for businesses and students/teachers. At H-FARM College, we're not just teaching AI; we're starting to prepare minds to use this powerful tool responsibly.  Businesses are rushing to adopt it, saving time and money and shifting workforce to perform more productive tasks in less time, similar to what happened when emails and the Internet appeared. This is going clearly to shift the balance as the next technological revolution and the best thing to do is help the new generations master the change, rather than falling behind. Looking ahead with optimism "If AI can do everything, what should I study?", a student asked me. We have some very strange beasts to deal with here and we need to make sure that the response and outputs in general is what we really want and that it is correct. To do this, we need to have real critical thinking skills to establish a successful collaboration with AI and arrive together at a result that is greater then the parts, AI + Human.

In this journey, I believe the OpenAI DevDay was a critical moment in reaching more accessible AIs that can be shaped by students, teachers, and everyone else. The potential for innovation is huge, yet we need to think about thoughtful application and ethical consideration. I can't wait to see what comes next! Are you passionate about these scenarios? Relive what we discussed at CLUB - Artificial Intelligence.