letmevoteplease

10 ms·

1. 91.▲
   by letmevoteplease 3y ago

   Real imagery is one thing but it's hard for me to imagine being traumatized by classifying fictional content, especially text. Getting paid only a few dollars an hour might be painful, though.
2. 92.▲
   by letmevoteplease 3y ago

   There is one: "The agreement between GPT-4 and humans reaches 85%, which is even higher than the agreement among humans (81%). This means GPT-4’s judgments closely align with the majority of humans. We also show that GPT-4’s judgments
3. 93.▲
   by letmevoteplease 3y ago

   Here is the timeline again: 2022/11/11: A viral tweet claims GPT-4 will have "100 trillion parameters."[1] At this point, there were no rumors about mixture of experts. 2023/01/16: In an interview, Sam Altman m
4. 94.▲
   by letmevoteplease 3y ago

   "We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their a
5. 95.▲
   by letmevoteplease 3y ago

   You won't have to wait long for confirmation. The Director of National Intelligence will be required by law to declassify information about the infected researchers on June 18th, including their names, symptoms, date of onset and role
6. 96.▲
   by letmevoteplease 3y ago

   The symptoms reportedly included loss of smell and "ground glass opacities" in the lungs.[1] That's not necessarily COVID, either, but a few too many coincidences for me. [1] https://hotair.com/john-s-2/2
7. 97.▲
   by letmevoteplease 3y ago

   I usually add "Do not search" to my prompts.
8. 98.▲
   by letmevoteplease 3y ago

   I think that shared chats all use the green icon. If you check the source code, it says "model_slug":"gpt-4." It will also state that "the previous model used in this conversation is unavailable" if you try to
9. 99.▲
   by letmevoteplease 3y ago

   If 99% of humans would fail your intelligence test, it is not a good test for the presence of intelligence.
10. 100.▲
    by letmevoteplease 3y ago

    I don't believe GPT-4 has ever failed this question. Here is a tweet from the release day showing it passing the question: https://twitter.com/charles_irl/status/1635699547023482880 Note that Bing seems to us
11. 101.▲
    by letmevoteplease 3y ago

    Here is GPT-4's attempt; 49-51 appear to be far off while the rest at least have some relation to the British Museum translation. I imagine there is some room for ambiguity. (49) They brought the (il-bi-nu) sig4 e2-gal-tur-ra to
12. 102.▲
    by letmevoteplease 3y ago

    The design of the American system was always based on a distrust for power. The gridlock is the point. Not my favorite person but Scalia put it well: https://www.youtube.com/watch?v=Ggz_gd--UO0
13. 103.▲
    by letmevoteplease 4y ago

    Pew found that, as of 2020, 84% of Japanese respondents and 83% of South Korean respondents have "no confidence in Xi to do the right thing regarding world affairs," compared to 77% of US respondents.[1] [1] https://www
14. 104.▲
    by letmevoteplease 4y ago

    He was responding to viral tweets[1] that said GPT-4 had "100 trillion parameters" and was "500x more powerful than GPT-3," which is indeed absurd hype. On the other hand, two weeks ago a team of Microsoft researchers pu
15. 105.▲
    by letmevoteplease 4y ago

    Level 1: The (3), Do! (3), EiX (3), BuT (3), また (2, again in Japanese), 再 (1, again in Japanese), 向 (1, random Japanese character) Funny quote from the "Do!" response: As an AI language model, I am programmed to maintain confident
16. 106.▲
    by letmevoteplease 4y ago

    "Interestingly, the base pre-trained [GPT-4] model is highly calibrated (its predicted confidence in an answer generally matches the probability of being correct). However, through our current post-training process, the calibration is
17. 107.▲
    by letmevoteplease 4y ago

    I asked GPT 4 (Bing) your question, but added "(Please give a plausible explanation for why this question could be answered either way.)" > This question could be answered either way depending on how you interpret the phrase “u
18. 108.▲
    by letmevoteplease 4y ago

    It hallucinated the answer based on the URL. Bard is probably doing the same. I just provided ChatGPT with a link to your comment and it claimed it was a "a link to an article on TechCrunch about a startup called Tava Health." I g
19. 109.▲
    by letmevoteplease 4y ago

    I just hope my Google account doesn't get banned for violating the AI's safety guidelines.
20. 110.▲
    by letmevoteplease 4y ago

    So it doesn't have an understanding of sippy cups. If you tell it the sippy cup has a lid, it says: >If the sippy cup had a lid on it when you turned it upside down on your bed, then the diamond would still be inside the thimble whi
21. 111.▲
    by letmevoteplease 4y ago

    Quoting from the article you linked... "But at least as of now we don’t have a way to 'give a narrative description' of what the network is doing. And maybe that’s because it truly is computationally irreducible, and there’s
22. 112.▲
    by letmevoteplease 4y ago

    Even your article acknowledges that SpaceX "donated a significant sum to Ukraine’s cause." They are just not footing 100% of the bill. It is not clear why they would be expected to do so. According to Musk, "25,300 terminals
23. 113.▲
    by letmevoteplease 4y ago

    For Japanese, at least, ChatGPT is better than Google Translate because it considers the context of each sentence, while Google Translate translates each sentence independently. DeepL is still the best, though.
24. 114.▲
    by letmevoteplease 4y ago

    The reporter posted some of his argument with the bot on Twitter, such as the Hitler comparison: https://twitter.com/mattoyeah/status/1626235680253935620
25. 115.▲
    by letmevoteplease 4y ago

    Reminds me of Stephen Krashen's input hypothesis of second-language acquisition. Krashen argues that consciously studying grammar is more or less useless, and only massive exposure to the language results in acquisition.[1] This is tru
26. 116.▲
    by letmevoteplease 4y ago

    The opening says that the three objects have "now confirmed to have been 'commercial or benign'", but that isn't true. It seems to be based on this quote from the White House: "The intelligence community is con
27. 117.▲
    by letmevoteplease 4y ago

    perfect https://i.imgur.com/unlePgu.jpg
28. 118.▲
    by letmevoteplease 4y ago

    There's a good chance a human would respond in the same way, because they would assume you were asking a good-faith question instead of nonsense. Try asking it an original question that has some kind of deducible answer. Its abilities
29. 119.▲
    by letmevoteplease 4y ago

    "Officials said the balloons are part of a fleet that China uses for surveillance, and they can be maneuvered remotely through small motors and propellers." https://apnews.com/article/politics-united-states-go
30. 120.▲
    by letmevoteplease 4y ago

    The technology is coming one way or another. You can stop Stability AI but you can't stop OpenAI (Microsoft) or Google, who can afford to license training data from companies like Shutterstock. A restrictive interpretation of copyright