Can ChatGPT o1 actually *think*?
Can OpenAI’s new GPT o1 models think and reason? What does it even mean for AI to think?
Subscribe for more tech commentary + jokes 🤪
References
“How many Rs are in the word strawberry” short: • Why no AI can solve this question
o1 model system card (aka the 43 page PDF): https://cdn.openai.com/o1-system-card...
o1 model dev docs: https://platform.openai.com/docs/guid...
00:00 OpenAI said what??
00:55 Computer can’t think
1:21 We’re going on a toaster metaphor
1:51 Anthropomorphizing AI
3:05 o1 + reasoning tokens
4:33 CoT prompting
5:18 The model is impressive … but not reasoning
6:14 Extremely important twitter lore (🍓)
8:00 Results of hype
8:40 No really, the model is impressive
9:25 Bonus trivia
Socials
/ alberta.tech
/ alberta.nyc
Set up
Breadboard: • Building an 8-bit breadboard computer!
Lego flowers: https://amzn.to/3XPMXFs
Lego bonsai: https://amzn.to/47xULit
Lighting: https://amzn.to/4gwvJV3 (and the good old fashioned sun)
#ai #tech #chatgpt #openai #softwareengineering
FTC Disclosure: This video contains affiliate links. If you purchase through one of these links, I may earn a small commission at no additional cost to you.