Last Edited on 5/13

Finally, the AI bit is picking up speed. OpenAI has just launched a set of videos demonstrating capabilities of its AI.

Since I’m a bit short on time and want to just put something out there, I have a couple of thoughts I’ll try to organize on my take of AI and its current state.

Without getting into the details of what exactly AI is, I guess I would sum up AI in its current state as the culmination of data aggregation and progressively low cost computing power/storage. With an extraordinary amount of data being ingested and processed in the last 20 years, and with computing power becoming more cost effective, we now have created something that resembles an incredibly enhanced version of “Google”.

  1. Human Thought and expression
  2. Real World Application
  3. Latency
  4. Uncanny Valley

Human Thought and expression

I hold true that crisp communication especially written communication is incredibly important to human thought and expression. With many example of AI’ified summaries, job postings, resumes, so much of the art of condensing information into something digestible is abstracted away. Is the human brain ready to read something that has been through the AI filter hundreds of times? A post that has been summarized, then that post summarized, then again and again? A game of AI telephone?

Real World Application

At Google I/O 18, there was a demo of a Google Assistant Haircut Appointment Call. It showcased a “realistic” interaction between someone booking an appointment and . I think in the last ~6 ish years, IVR (Interactive Voice Recordings) have become more prevalent, as well as online chat bots as well as self-service web applications.

It’s rare for me to have witnessed an actual real world application that I find particularly impressive (although a colleague recently shared that ChatGPT was able to identify a specific piece of code that caused a security related bug)

I think there are some interesting applications of AI, and that they reside in much less flashy stuff that’s being shown to the public these days, and much more monotonous things like….summarization.

Latency

OpenAI’s suite of Youtube videos demonstrate a wide array of activities. Whether it be my ADD brain or what, I find the videos don’t seem to capture the concept of latency very well. In computing, latency is always some factor, and presumably, for AI to effective, it has to be connected to some datasource somewhere.

Where in the hell are people sustaining the type of internet connection to be able to reliably generate such? I get frustrated daily with having to restart apps, with websites being overloaded with ads on my mobile phone, and with my entire experience on the internet being inundated with advertisement after advertisement.

Uncanny Valley

Of course, there’s the ultimate reminder that the difference between AI and us is the human element. Reading something out of ChatGPT versus something that I would describe as “well-written” doesn’t really pair well. It gets the job done, but in the same way that AI produced artwork, and even AI produced music (that’s gotta be a thing) is definitely in its own genre. The human population is already gearing up to now add another question to their reality of “is this AI generated?” , adding onto the already long list of concerns humans have for their realities.

Really though, this is only my initial impression of what’s been going on these last couple of months. I’d like to think I’m open minded about what the future of AI looks like and what innovation it brings. I do however think we are in the “gold rush” period of any new type of Tech from Silicon Valley, and only in the end will the people with a strong conviction of their ideas be able to push the applications of “AI’”

Reminder for movies I need to watch – AI Related

Movies I need to watch

  • Terminator Series
  • The Matrix (I need to watch the whole series I suppose)