dynomight

Scattered thoughts on how doing math feels these days dynomight.net/math/

me: what kind of idiot has their refrigerators connected to the internet? them:

has anyone else noticed this

Strongly feel that the ultimate destination of this kind of argument is that consciousness does not come in discrete clumps.

Grace@gracekind.net

If you believe that consciousness exists or occurs in a physical location, and that LLMs are conscious, you need to be okay with the idea of that location jumping around / being spread out over large distances

lemmy.world/post/5232721...

hmmmmm

What I'd like is for AI-risk skeptics to commit to publicly acknowledging they were wrong if they turn out to have been wrong, in the way that 90% of the people who supported Iraq War II never did.

teleology is good, i will fight you

The question you all were wondering: dynomight.net/liver/

13 Aug

hot take but walking around in public with thousands of strangers and ~zero concern anyone will stab you and take your stuff

norvid_studies@norvid-studies.bsky.social

what else is like this

11 Aug

Thoughts I had while reading your comment dynomight.net/comment/

4 Aug

Once found myself thinking: "Once I get these new self-watering pots, my life is going to really come together."

Tyler Alterman@tyleralterman.bsky.social

I’m terrified by my tendency to get swallowed up by my own stories: “This is the right moral philosophy.” “This person will save me.” “This project is what I should dedicate my life to.” 2yrs later: “ugh wtf” What are healthy ways to interact with this tendency?

3 Aug

This post went in a weird direction. dynomight.net/human-alignm...

31 July

What I really find surprising is that there's a huge demand for AI writing detectors but (seemingly?) almost no demand for AI image/video detectors. I *definitely* would not have predicted that.

norvid_studies@norvid-studies.bsky.social

I think it's interesting and slightly surprising the degree to which AI image and video generation has been a "nothing happens" effect, compared to what one might have reasonably predicted in advance. now this may be my anecdotal experience firewalled off from "the bad internet",

So you want to use plants to reduce CO₂ dynomight.net/plants/

Thread of ideas that don't have short names but deserve them: 1. For most common issues, if you just research and follow standard boring conventional best-practices, you'll do better than ~95% of people, because they don't do that.

4. It's actually quite hard to find generalists who are simultaneously pretty competent and reliable and easy to work with. You get get surprisingly far in the world by just being mildly above-average on all those three axes at the same time.

5. This RCT showing a small but non-significant effect is completely consistent with a true small effect, or no effect.

3. Sometimes when people say they aren't worried about something but give weird and shifting explanations, that's because they don't actually believe that thing is possible, but haven't fully acknowledged that fact to themselves.

(I stress that these ideas are not necessarily correct, but if nothing else we could shorten debates by having them once and for all and giving them labels.)

2. AI is exhausting and many of us are sick to death of thinking about it, but it's so massively consequential (potential AGI, obviously, but even just to the short-term US economy) that not thinking about it is not an option.

28 July

for posterity: commentator accused of being an AI agent immediately confirms being an AI agent

28 July

On a near-daily basis I get emails offering to pay for "sponsored articles" or for linking to sketchy crypto/casino sites. I guess everyone gets these? Search engines surely prohibit this, but does it get your content removed from LLM training?

28 July

Does every question mark deserve a Betteridge? dynomight.net/betteridge/

22 July

Does creatine make you smarter? dynomight.net/creatine/

21 July

Austin Vernon on one of the more frightening uses of market mechanisms: www.austinvernon.site/blog/ukraine...

21 July

precedent

dynomight@dynomight.net

Pseudpocalypse dynomight.net/pseudpocalyp...

16 July

this is one possible way to visualize information

15 July

Whichever scraper is going through my blogposts, one-by-one requesting 50-100 copies of each at the same time, once per hour, while pretending to be referred from Google and running javascript and not blocking goatcounter: Why are you doing that?

14 July

Pseudpocalypse dynomight.net/pseudpocalyp...

6 July

Here's a post on hazard ratios, which I expect to be one of the least popular things I've ever written. ♡♡♡ dynomight.net/hazard-ratios/

29 June

how it started // how it's going

26 June

Blink if you're human dynomight.net/blink/

23 June

The worthlessness of vitamin D is mildly exaggerated dynomight.net/vitamin-d/

11 June

I need blog recommendations, please send blog recommendations. (Do not let modesty prevent you from recommending your own blog.)

11 June

(ideally blogs that are ≤ 90% about AI but whatevs)

2 June

The fact that humans have different emotional reactions when witnessing an event versus hearing about it from a credible source is surprisingly central to explaining the state of the world

2 June

Counterpoint: Perhaps this is just because we become inured to anything we experience a lot? If witnessing was common and hearing rare, maybe it would be the opposite?

31 May

Starting to suspect that the same qualities that would make someone flee twitter also tend to make you hesitant to engage too much with microblogging (or internet life) in general.

visa@visakanv.com

does it feel like bsky still gets new waves of people fleeing X, or does that feel like it’s been done for a while?

26 May

If you're a blogger (or aspiring blogger) interested in "progress", I recommend you consider applying for the roots in progress blog-building initiative. (Deadline June 1) rootsofprogress.org/fellowship/

WHAT IS A SIMCLUSTER

Is "colorectal cancer" rising in "young people"? dynomight.net/crc-rates/