🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

we often disagree with Yann leCun, but he might be right about this: "human-level AI is still at least one major breakthrough away, if not several" the big question is: when will that breakthrough come?

in the next 3–5 years, we’ll need to see AI applied in the real world before we can confirm AGI AI already knows more than i do, but it still lacks judgment and agency the agency is being worked on now, and judgment will come from more real-world experience to get there, AI

Terence Tao says a major leap this decade could be AI forming deep, original conjectures -- linking ideas no one thought were connected that dream extends to physics: discovering new laws just from raw data but current models can’t even rediscover old laws reliably, often

can AI help to cure all diseases? when i talk about AI curing all major diseases within a decade, people look at me like i'm crazy and honestly, it is a wild claim so if you want people to see AI as helpful, show real examples of what it’s already done ...not just big

"AI can do your job" sounds cool… but it hits differently when you realize it is doing your work they say: "don’t worry, new jobs will come" but ask what jobs? you get hand-waves and hypotheticals the truth? we’re already seeing the shift, and it’s happening faster than

what sam and demis call AGI would be considered ASI by most standards -- an AI as smart as a team of experts and that might be possible. but ASI, as many imagine it, isn’t. it still runs on a turing machine, so its thinking isn’t beyond us ...and it can’t do the impossible,

i believe all the easy wins are gone things aren’t easy anymore, and openAI won’t have a one-year lead, if they ever get a lead again but ppl still use chatgpt a lot, by a huge margin, and it keeps growing fast sam knows how things work -- he helped spread the idea of fast

still blows my mind that most ppl i know barely use AI, maybe for around 5-10% of their tasks i use AI for 70-80% of my work, and have used it for the past two years it either helps a bit or fully handles tasks like: • coding, bug fixing • summarizing research • building

o3-pro does feel different i used it to help on a small business strategy document this week. i expected outputs similar to what i get from gemini 2.5 pro and claude 4 opus but o3-pro seemed to take my fairly light prompt and run deep with it the quality of the output took me

Geoffrey Hinton says AI agents may need emotions to be effective in human roles For example, In a call center, if someone wants to chat all day -- the AI agent acts bored or irritated to end the call "if it behaves like it’s annoyed, i’d say it has an emotion"

i honestly don’t get how ppl still deny gpt-4 as some kind of narrow AGI yes you can highlight its limitations all day, ...but we're looking at a system that understands plain human language and delivers intelligent responses on nearly any topic we’ll see more upgrades and big

AI is either dismissed as hype or reduced to novelty but both views miss the point. what we're witnessing isn’t a trend -- it's a foundational shift in how intelligence is built, utilized, and distributed the smart ones don’t argue if AI is "cool" they’re preparing for what

i'm wondering which jobs AI can't replace? even if we had ASI and perfect robots that could do everything better than us ...wouldn't there still be some jobs that feel more valuable when done by humans? i think so... not because we need to, but because we want to