Correction for the record: METR is not funding my work directly. I am participating in a METR study regarding performance effects of software development AI-assistants & have chosen the project above to be my primary focus during the study.
To avoid hitting HN's rate limit, I will be adding my answers to this post for as long as I am able to edit (2 hours, I think). Once I am no longer able to edit or have hit a character length restriction, I will create a reply post to this one and do the same thing there.
Also, to reduce confusion, the original title was "Show HN: A belief system is how we create AI that actually think like humans do"
Answer to taylodl's question[0]:
AI will never be able to think exactly like humans do. There will be trade-offs to this. However, enabling AI to think more like humans will enable both AI & humanity to better explore these tradeoffs.
Reply to Voland0's post[1]:
I've found trying to specify my own personal solution process to be very enlightening, but also quite mentally taxing. Such self-awareness comes at quite the cognitive cost!
> How is GP's idea related to 3~4 different home projects of yours and not just one?
My thought process is that if I actually make this for myself, those 3-4 projects would magically get a very impatient new stakeholder who will pester me to actually deliver those projects to them.
It's gonna be interesting to see if I'm able to trick my brain into not just brushing off the agent (or if that's even really an issue, no clue at this point). I've started rolling around some ideas on handling that scenario but I'm just gonna let that stew while I play with different setups so I don't end up just building a convoluted reminder app lol
Unfortunately I had an unexpectedly hectic holiday weekend so I won't get to start playing with the idea for real until tomorrow afternoon.
Personally, I think we need much tighter regulation. There was a study at some point that found that lobyying has a yield of 22,000% in the US, which is a wild figure (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1375082). Lobbying very seriously undermines democracy. The US system of calling lobbying via donations as free speech seems ludicrous. France, for instance, has much tighter laws on lobbying.
I think marrying someone despite being attracted to the other sex without telling them and then having an affair with someone definitely makes you a bad person. But that’s me.
I can even tolerate / excuse / forgive up until that point, because it is indeed an unfair system. She took a gamble and got caught, at which point she ought to have made a deal with the guy. Not exploited the other unfair system of state violence against him.
Not op, but markdown is much more likely to render well in different contexts, without post processing. My editor understands markdown, GitHub understands markdown, the link preview renderer in <random collaborative tool> understands markdown. It’s the lowest common denominator
That's true, and it's why we're all using it. But those different renderers all support different ill-defined interpretations of Markdown. You can forget about all of them accepting raw HTML.
It has sufficient differences to what is already accepted "everywhere" that I would have think about syntax more often than I'd like. That is enough. The minor inconveniences of Markdown incompatibilities are smaller than the inconveniece of AsciiDoc. It simply doesn't offer nearly enough potential advantages to be worth the hassle.
reply