energy123

42 ms·

1. 1021.▲
   by energy123 1y ago

   If investment was the only major issue, rental inflation would not be high. Rental inflation can only be explained by a lack of housing in areas people want to live.
2. 1022.▲
   by energy123 1y ago

   When I look at disciplines other than my own, I can recognize the skill and elegance that goes into it, even though I don't have those skills myself. But when I look at the output of web development, and I see smooth scrolling, scrollj
3. 1023.▲
   by energy123 1y ago

   The general framework would be to sort by `U(dist(ratio))`. The choice of `U` (utility function) being a separate question to the estimation of `dist(ratio)`.
4. 1024.▲
   by energy123 1y ago

   If you remove bad regulations and enhance the good regulations, then yes.
5. 1025.▲
   by energy123 1y ago

   > new GPT Another sign tapping is needed. > AI is exceptional for coding! [high-compute scaffold around multiple instances / undisclosed IOI model / AlphaEvolve] > AI is awesome for coding! [Gpt-5 Pro] > AI is somewhat
6. 1026.▲
   by energy123 1y ago

   Basic prose is a saturated bench. You can't go above 100% so by definition progress will stall on such benchmarks.
7. 1027.▲
   by energy123 1y ago

   Basic prose is a saturated bench. You can't go above 100% so by definition progress will stall on such benchmarks.
8. 1028.▲
   by energy123 1y ago

   They could have taken the enriched uranium out and used it to shorten the critical path, no?
9. 1029.▲
   by energy123 1y ago

   How tightly scaffolded/harnessed/constrained is your primary agent for a given task? Are you telling it what reasoning strategy to use?
10. 1030.▲
    by energy123 1y ago

    I have to tap the sign whenever someone talks about "GPT-5" > AI is exceptional for coding! [high-compute scaffold around multiple instances / undisclosed IOI model / AlphaEvolve] > AI is awesome for coding! [Gpt-5
11. 1031.▲
    by energy123 1y ago

    Have you tried greedy decoding (temp 0) in aistudio? The temp 0.7-1.0 defaults are not designed for reconstructing context with perfect accuracy.
12. 1032.▲
    by energy123 1y ago

    You aren't using those models. They aren't released.
13. 1033.▲
    by energy123 1y ago

    > "All these developers are building is just expensive new luxury apartments" I've recently realized that any argument that contains the word "just" can be discarded.
14. 1034.▲
    by energy123 1y ago

    It's an explanation, as much as you wish it isn't.
15. 1035.▲
    by energy123 1y ago

    It's strange seeing so many takes like this two weeks after LLMs won gold medals at IMO and IOI. The cognitive dissonance is going to be wild when it all comes to a head in two years.
16. 1036.▲
    by energy123 1y ago

    It's an explanation not a justification.
17. 1037.▲
    by energy123 1y ago

    I can't agree there's a plateau just a few weeks after two companies got gold medals at IOI and IMO using natural language (no Lean). Seems like progress is continuing nicely.
18. 1038.▲
    by energy123 1y ago

    You either want a plan like China or an absence of planning like Texas. Either help like China or get out of the way like Texas. Two places that can actually build energy. Don't be like California where the government doesn't help
19. 1039.▲
    by energy123 1y ago

    It's the same. Every assumption, taste, doc, opinion, edge case, unstated knowledge, requirement, exception to the requirement, intention, should be provided for best results. It's not crucial for throwaway weekend projects, but f
20. 1040.▲
    by energy123 1y ago

    I think it's a lot more nuanced than this. Defense can be perceived as threatening and cause others to increase their defense, as per Balance of Threat theory and the security dilemma, creating a feedback loop (arms race) that leads to
21. 1041.▲
    by energy123 1y ago

    > Without Hamas, there can be no Natanyahu. Without the Second Intifada, you mean. That's the moment when Israel's political left was obliterated according to domestic polling, and it hasn't yet recovered.
22. 1042.▲
    by energy123 1y ago

    > AI is exceptional for coding! [high-compute scaffold around multiple instances / undisclosed IOI model / AlphaEvolve] > AI is awesome for coding! [Gpt-5 Pro] > AI is somewhat awesome for coding! ["gpt-5" with
23. 1043.▲
    by energy123 1y ago

    Greedy decoding isn't random.
24. 1044.▲
    by energy123 1y ago

    Morality is self-interest. Piracy isn't immoral because they aren't staked in the impacted industries. Eating meat isn't immoral because it's in their self-interest to keep doing it. What is immoral are things that wou
25. 1045.▲
    by energy123 1y ago

    Performance is proportional to the number of reasoning tokens. How to reconcile that with your opinion that they are "random words"?
26. 1046.▲
    GPT-5 Reasoning Effort (Juice): How much is used in the API vs ChatGPT(old.reddit.com)

    2 points by energy123 1y ago|0 comments
27. 1047.▲
    by energy123 1y ago

    > My personal test question keeps bombing, and I think it's something they should be capable of doing? The problem is that "they" isn't a monolith. How much compute went into your tests? Gpt-5 thinking in ChatGPT Plus
28. 1048.▲
    by energy123 1y ago

    How can you say progress has stalled two weeks after LLMs won gold medals at IOI and IMO? How can you say progress has stalled without having visibility on the compute costs of gpt-5 relative to o3? How can you say progress has stalled by r
29. 1049.▲
    by energy123 1y ago

    Good job being explicit about the reasoning effort and verbosity settings. And interesting but not surprising that verbosity helped performance.
30. 1050.▲
    by energy123 1y ago

    It's not a question of whether it's the case. It's confirmed by OpenAI employees on Twitter. The reasons could be that it's new (they did say they plan to release eventually but not soon), or that it's too heavily s