Project ideas from Hacker News discussions.

The changing goalposts of AGI and timelines

📝 Discussion Summary (Click to expand)

Five key themes that dominate the discussion

# Theme Representative quotes
1 Ethics of AI in defense & surveillance I have deep respect for Sam and the team, and I’m proud of what we built together… this was about principle, not people.” – Caitlin Kalinowski
Why is it so business as usual for them? Do they consider this at par with missiles with infra‑red/heat sensors for tracking/locking?” – sheepscreek
2 Corporate politics & whistle‑blowing OpenAI will never do this.” – bilekas (referring to the Pentagon deal)
If you sell a weapon to the department that is in charge of killing people… you don’t get a say in who gets killed or how.” – lich_king
3 Debate over AGI definition & timelines We’re not near AGI yet, but we’re close to it in some domains.” – bigyabai
AGI is a type of artificial intelligence that matches or surpasses human capabilities across virtually all cognitive tasks.” – OpenAI charter
4 Technical limits of current LLMs They are just next‑token predictors; that’s why they can’t truly understand.” – sulam
They can predict the next token, but that doesn’t mean they have a real‑world model or learning ability.” – sulam
5 Supply‑chain risk & government contracts They want to subsume the work you’re already doing with more extreme limitations… it would be really convenient if OpenAI just agreed to terms.” – bigyabai
Labeling them a supply chain risk is an extreme response.” – BoxFour

These five themes capture the bulk of the conversation: the moral stakes of AI in warfare, the internal politics of the leading companies, the fuzzy nature of “AGI”, the concrete shortcomings of today’s models, and the commercial‑government dynamics that shape the industry.


🚀 Project Ideas

Error generating project ideas: 'choices' - Monetization: Hobby

Read Later