Today I met someone interning at AT&T, a major US company providing telecommunications services (though I should probably double-check that!). He mentioned he joined the company as a newbie developer. I need to ask him more about this, but after he got there, they formed a small team he became a part of. I'm not entirely sure how. This team specifically focused on implementing and integrating AI into the company – figuring out how and where they could use AI to boost efficiency. It seems they're really keen on updating their systems and incorporating AI. But here's the thing: there aren't many people who truly understand AI's potential. And those who know how and where to implement AI are super valuable, in my opinion. Staying on top of AI trends, tools, and everything AI-related is an important skill. And then, connecting that knowledge to existing problems. It reminds me of what businessmen like Steve Jobs and Deep Maheshwari said about "connecting the dots," linking solutions to problems. That, in itself, is entrepreneurship.
Also, I heard from OpenAI researchers and AI/ML engineers that building a GPT-4 model from scratch might only require 5-10 people 😯 (mind-blowing!).
This is what they're saying about their next plans, or what they might have already started working on. I'm not sure:
What’s next?
- Smarter use of data: Humans learn from just a few examples. AI doesn’t — yet. Getting AI to learn more like humans is the next big step.
- System upgrades: The future involves making more fault-tolerant systems so things don’t crash often.
- 10 million GPU runs? They think it's possible someday, but it’ll look different — maybe decentralized or semi-synchronous.
I'm sharing some things that felt important and helpful in understanding where things are heading.
