A great conversation about neural networks with Igor Kotenkov, ML engineer and creator of the "Seeallochnaya" channel.
Igor's resources:
Telegram: https://t.me/seeallochnaya
LinkedIn: / seeall
Lecture on Transformers (for those who aren't afraid to delve into it): • Transformer, explained in detail | Igor Ko... - subscribe, he promises lots of content...soon. He even bought a microphone!
All of Igor's materials can be found on Telegram: https://t.me/seeallochnaya/3
And, if you haven't read them yet, I recommend three articles co-authored with Pavel Komarovsky and other authors:
1. https://habr.com/ru/company/ods/blog/... — about the structure of ChatGPT
2. https://habr.com/ru/company/ods/blog/... — what's different about the recently released GPT-4
3. https://vas3k.blog/blog/ai_alignment/ — about the risks posed by the emergence of AGI
My social networks:
Support the channel on Boosty: https://boosty.to/digital_ninja
Telegram: https://t.me/digital_ninjaa
Live Telegram channel: https://t.me/sanya_ilyin
Twitter: / digital_ninjaaa
Instagram: / jjoskey
00:00:00 - Introduction
00:00:17 - Is ChatGPT a Technological Shift?
00:02:54 - Attitudes to the Phrase "Artificial Intelligence" in the ML Community
00:04:28 - What is AGI and When Will It Appear
00:12:00 - Society After AGI
00:24:04 - How to Bypass ChatGPT Limitations
00:29:24 - Why OpenAI Is Closing Information About GPT-4
00:35:11 - What is a Token and Why is It Needed
00:44:47 - How Side Effects Occur in a Model
00:54:11 - Model Limitations
00:57:30 - What Are the Uses Did the cases surprise Igor?
01:05:01 - What will happen to programmers?
01:14:27 - AI Alignment: what is it and why is it needed?
01:17:15 - People are panicking over the emergence of AGI
01:20:01 - Igor's Career