Sparks of AGI

So our quest for artificial superintelligence might be complete much faster than we thought.

Microsoft recently released a paper titled “Sparks of AIG”.

They claim that OpenAI is showing that it “could reasonably be viewed as an early version of an artificial general intelligence (AGI)”

An idea that was mocked by AI researches just 5 to 7 years ago is now published as research by Microsoft.

Here is me reading over the research paper and explaining what it means:

(this is a big deal folks)

Whatever timeline we though we were on, whatever time we though we had to get ready is rapidly shrinking.

Since December 2022, this system went from being a somewhat error-prone high school student to now beating college graduates and even PhD level experts at exams, coding and complex reasoning.

The latest upgrades shows that it can use tools in areas where it’s lacking.

It calls a calculator function for math problems it can’t solve.

It runs an internet search to get the latest data it doesn’t know.

It repeatedly shows the ability to solve problems that it wasn’t trained to do (aka general intelligence)

…but it’s not all sunshine and rainbows.

There are some very real and very valid concerns over the fact that Microsoft and OpenAI now possibly control this tech without ANY oversight or visibility into what they are doing.

This video is long (30 mins), but I added chapters that highlight the most important parts.

I hope you take a second to watch it at least in part.

This thing is coming fast and most people are asleep at the wheel…

Your Truly,

Wes “Don’t Panic” Roth

PS I made a video about the emergence of AGI. If you are in hurry at least watch the first few mins to understand why this is a game changer: