Thought Banana
What DeFi Wishes It Was
At this point, we could do an AI in every section just about every day. Hell, might as well even have ChatGPT write it too. Actually, wait, forget I said that…and please, nobody give Patrick that idea either.
Ahem, anyway. The world of artificial intelligence continues to evolve rapidly. While the evolving done by digital currencies throughout their own hype cycle was one of price movements, laser eyes, and “have fun staying poor,” the evolution of AI seems to have already gone from idea to countless real-world use cases in the same amount of time BTC went from $0.0008 to $0.10.
Quite honestly, there are few better places to find these updates than Twitter. If you search “AI” on the platform, odds are you’ll stumble across something wild I missed, as the sector is quite literally oozing with new updates.
For starters, ChatGPT is now available as an iOS app, newly released as of May 18th. Accessing the GPT-4 LLM is now just one click away every single time you open up your phone…and to think, only 6 months ago, you were still using Safari.
Earlier in the week, a company called Gamma released an app that just may save all you IB hardos 10 of the 90 hours you work each week by creating slide decks for you simply based on the natural-language words you use to describe what you want. And yes, it’s free, but no promises on whether or not your MD will catch you.
Meanwhile, for all you ER and other analysts out there, speed reading industry reports gets taken to the next level with ChatPDF. Like a gift from the heavens, you can drop any PDF file at all into this platform, and it will immediately become, effectively, an LLM specialized in that one PDF. You can then ask the bot any questions about the PDF, and it doesn’t take much imagination for where things can go from there.
Also released on May 18th was Perplexity AI’s newest update to its CoPilot search tool, allowing interactive and personalized search designed to save time when browsing.
But while most companies run into the AI battlefield ready for action, one particularly notable is moving in the opposite direction. Last week, Apple banned certain employees from using these LLMs as the firm is wary of leaks related to product design, financial information, or other sensitive data that would cause a historically secretive company like Apple to throw up if leaked.
At its most fundamental, what these LLMs allow us to do is interface with a computer using human language. Previously, that interface was reserved to nerds speaking C++, Javascript, or some other bullsh*t. But taking things one step further, the LLMs essentially allow you to interface with the entirety of the internet in one place; rather than interfacing with whatever link pops up first on your Google search, you’re able to interact with the internet’s hivemind in its entirety all in one place.
This “bundling” of essentially the entire internet in one place, needless to say, has countless possibilities on its own, with just a few pointed out above. As the hardware behind these things, like Meta’s new AI chips and others, continues to get better, the possibilities will only mount.
There’ll still be a massive bubble, I’m sure, but very unlike BTC and digital “assets,” this one actually has a consumer-facing use case. The internet revolution gave us the dot-com bubble, where countless companies listed publicly and ran up to multi-billion valuations just for having “.com” in their name. Out of that, you got a ton of Pets.com meltdowns, but you also got Amazon and Google.
Lesson of the day? Choose wisely.
The big question: What even are the appropriate questions to be asking about the long-term development of AI? What are some of the greatest economic opportunities with this technology? And, of course, how close are we to Skynet?
|