4 Comments

That is a manifesto if there ever was one:

• This is Important.

• We are Ready

• Art Matters.

AMEN! I joyfully and whole-heartedly add my Name to the signatories.

Thank you sincerely so much for so beautifully articulating this present moment, doing it with style, class and hope, and pointing out a shining silver lining on the brewing thunderhead.

Expand full comment
Comment deleted
Expand full comment

...and then, within 24 hours of each other, MidJourney 5 and GPT 4 were released on the world. It is most definitely... here.

Expand full comment
Comment deleted
Expand full comment

But, Mitchell, they don't *need* to "impress" you. They merely need to make a significant impact on humanity, industry, and perhaps even culture and civilization. Plus, as the original poster notes... this is the very very very beginning. This is the first "spark". Fire. fire > electricity > combustion engines > nuclear reactors > fusion. The difference is, that process took 100s of thousands of years. This one might take 100s of *days*

Expand full comment
Comment deleted
Expand full comment

Agree. GPT-3 gave us a gesture, a hint, a tease, a foreshadow. This GPT-4 tho, imho it has broken through the "glass ceiling" of AI. It now sits firmly within the bounds of HLMI. we can argue whether its IQ 80 or IQ 130 (its certainly "uneven" competence across task areas / specializations), but again, doesn't matter. Its in the range. Its smarter than some humans. It has *already* begun replacing human jobs... knowledge workers! white collar! So it has the "foothold". and it doesn't go "backwards" from here. Today, it can write functional, solid code to control my robot arm and read data from sensors. A primitive python script (that it also writes) can pump those results back into GPT. Its a kludge today, but with proper API access and speed... its a stone's throw away from real-world experience and actuators. And... no-one really has any idea what GPT4 is even capable of. We won't know until we poke and prod and test and iterate.

Expand full comment