4 ms·

what consensus? i think most researchers remain skeptical

bpodgursky 4y ago
Uh... no. Most researchers have moved their timelines to somewhere between 2030 and 2040. You can argue they're wrong, but there is absolutely a general consensus that AGI is going to be this generation.
- xen2xen1 4y ago
  And consensus is never wrong!
  - SketchySeaBeast 4y ago
    Especially assertions of consensus provided without evidence of said consensus.
    - gfodor 4y ago
      I don’t know that the consensus is right but the OP’s confidence of it being far off made me think they were unaware of the recent shifts (not just in consensus, but the upstream capabilities.)
- gizajob 4y ago
  AGI has been 20-30 years away for some 70 years now...
  - Isinlor 4y ago
    Kurzweil in 2002 made $20,000 bet that a difficult, well defined 2h version of Turing test will by passed by 2029. https://longbets.org/1/ https://longbets.org/1/ Given development in language models in the last 2 years he may have a decent chance at winning that bet. People give him 65% chance [0] and by now there are only 7 years left. [0] https://www.metaculus.com/questions/3648/computer-passes-turing-test-by-2029/ https://www.metaculus.com/questions/3648/computer-passes-tur...
- _delirium 4y ago
  Who do you have in mind? In my corner of AI it's pretty uncommon for researchers to even predict "timelines". Predictions have a bad track record in the field and most researchers know it, so don't like to go on record making them. The only prominent AI researcher I know who has made a bunch of predictions with dates is Rodney Brooks [1], and he puts even dog-level general intelligence as "not earlier than 2048". I imagine folks like LeCun or Hinton are more optimistic, but as far as I'm aware they haven't wanted to make specific predictions with dates like that (and LeCun doesn't like the term "AGI", because he doesn't think "general intelligence" exists even for humans). [1] https://rodneybrooks.com/my-dated-predictions/ https://rodneybrooks.com/my-dated-predictions/
- TaupeRanger 4y ago
  Sure...just like there was during the last episode of AI hype a generation ago.
semi-extrinsic 4y ago
Yeah, I don't think there is even any agreement about what criteria a "minimal AGI" would need to meet. If we can't even define what the thing is, saying we'll have it within ten years is pure hubris.
Isinlor 4y ago
The survey [0], fielded in late 2019 (before GPT-3, Chinchilla, Flamingo, PaLM, Codex, Dall-E, Minerva etc.), elicited forecasts for near-term AI development milestones and high- or human-level machine intelligence, defined as when machines are able to accomplish every or almost every task humans are able to do currently. They sample 296 researchers who presented at two important AI/ML conferences ICML and NeurIPS. Results from their 2019 survey show that, in aggregate, AI/ML researchers surveyed placed a 50% likelihood of human-level machine intelligence being achieved by 2060. The results show researchers newly contacted in 2019 expressed similar beliefs about the progress of advanced AI as respondents in the Grace et al. (2018) survey. [0] https://arxiv.org/abs/2206.04132 https://arxiv.org/abs/2206.04132

Uh... no. Most researchers have moved their timelines to somewhere between 2030 and 2040. You can argue they're wrong, but there is absolutely a general consensus that AGI is going to be this generation.

xen2xen1 4y ago
And consensus is never wrong!
- SketchySeaBeast 4y ago
  Especially assertions of consensus provided without evidence of said consensus.
  - gfodor 4y ago
    I don’t know that the consensus is right but the OP’s confidence of it being far off made me think they were unaware of the recent shifts (not just in consensus, but the upstream capabilities.)
gizajob 4y ago
AGI has been 20-30 years away for some 70 years now...
- Isinlor 4y ago
  Kurzweil in 2002 made $20,000 bet that a difficult, well defined 2h version of Turing test will by passed by 2029. https://longbets.org/1/ https://longbets.org/1/ Given development in language models in the last 2 years he may have a decent chance at winning that bet. People give him 65% chance [0] and by now there are only 7 years left. [0] https://www.metaculus.com/questions/3648/computer-passes-turing-test-by-2029/ https://www.metaculus.com/questions/3648/computer-passes-tur...
_delirium 4y ago
Who do you have in mind? In my corner of AI it's pretty uncommon for researchers to even predict "timelines". Predictions have a bad track record in the field and most researchers know it, so don't like to go on record making them. The only prominent AI researcher I know who has made a bunch of predictions with dates is Rodney Brooks [1], and he puts even dog-level general intelligence as "not earlier than 2048". I imagine folks like LeCun or Hinton are more optimistic, but as far as I'm aware they haven't wanted to make specific predictions with dates like that (and LeCun doesn't like the term "AGI", because he doesn't think "general intelligence" exists even for humans). [1] https://rodneybrooks.com/my-dated-predictions/ https://rodneybrooks.com/my-dated-predictions/
TaupeRanger 4y ago
Sure...just like there was during the last episode of AI hype a generation ago.

semi-extrinsic 4y ago

Yeah, I don't think there is even any agreement about what criteria a "minimal AGI" would need to meet. If we can't even define what the thing is, saying we'll have it within ten years is pure hubris.

Isinlor 4y ago

The survey [0], fielded in late 2019 (before GPT-3, Chinchilla, Flamingo, PaLM, Codex, Dall-E, Minerva etc.), elicited forecasts for near-term AI development milestones and high- or human-level machine intelligence, defined as when machines are able to accomplish every or almost every task humans are able to do currently. They sample 296 researchers who presented at two important AI/ML conferences ICML and NeurIPS. Results from their 2019 survey show that, in aggregate, AI/ML researchers surveyed placed a 50% likelihood of human-level machine intelligence being achieved by 2060. The results show researchers newly contacted in 2019 expressed similar beliefs about the progress of advanced AI as respondents in the Grace et al. (2018) survey. [0] https://arxiv.org/abs/2206.04132 https://arxiv.org/abs/2206.04132