Advertisement · 728 × 90

Posts by Josef Fruehwald

i love living in the first village of an rpg man. friendly neighbors, music is chill af, we even got that one kid we adopted into the village of whom the ancient prophecy was written about. anyway there’s a big ceremony tomorrow can't wait to still be alive after it

7 hours ago 1404 534 10 9

also interesting that the “how would you enforce it?” asker quickly pivoted to “LLMs are good actually.”

4 hours ago 1 0 0 0
Preview
Don’t seek permission, center values When you're enamoured of a technology and someone points out important ethical challenges, a typical reflex is to seek permission:…

It's okay, you don't need my permission ideophone.org/dont-seek-pe...

As for what I do or do not realize about LLMs, feel free to peruse my work doi.org/10.5281/zeno... or that of colleagues like www.sciencedirect.com/science/arti...

21 hours ago 31 5 2 0

just ocurred to me you could model any ratio with a log-link and an offset

brm(
bill_len ~ species + offset(log(bill_dep)),
data = penguins,
family = gaussian(link = "log")
)

7 hours ago 1 0 0 0

a DMing thing i’ve started doing: I realized i don’t like playing D&D against myself, so when there are friendly NPCs in the party and combat begins, they’ll call out “what should I
do?” and whichever PC answers, i hand over the stat block for them to run

1 day ago 7 0 0 0

has anyone ever *not* had divergent transitions on a real model with the default adapt_delta?

1 day ago 0 0 0 0

i was gonna joke, but telling people that english has more than 5 vowels (and they're not AEIOU) is shocking enough to some students that they never incorporate it into their knowledge of the world...

1 day ago 9 0 1 0
Advertisement

when getting predictions from brms models, does `re_formula = NA` also remove random-factor smooths?

3 days ago 4 0 3 1

Corresponding with the authors? Just straight to submitting a response manuscript?

4 days ago 1 0 2 0

ok, an ettiquite q: in analyzing a data set on OSF, i’ve found that a poppy result is probably much less reliable than reported in a paper (absolutely no indication of fraud or statistical malpractice). I think it’s maybe worth writing up *why*, but what’s the best first step?

4 days ago 2 0 1 0
Post image

acquiring multidialectal competence sets you up for surprise humorous moments no one else around you will get

4 days ago 6 1 0 0

Next they’ll partner with Charmin for the people who need to ask a chatbot how to wipe their ass

5 days ago 5 1 0 0

"The Company will initially seek to acquire high-performance, low-latency AI compute hardware"

hell yeah. Tangible goods and domain experience are out. Intangible promises with "AI" stamped on them are in!

6 days ago 1 0 0 0

is it a sign of a rational market when a shoe maker says they're AI now and their stock increases (at time of writing) 700%?

6 days ago 6 1 2 0
Advertisement

At this rate, we’ll stop speaking altogether by 2050!

(… this has the whiff of statistical anomaly about it, but I’ll have to dig into it later)

1 week ago 17 1 4 0

Three years ago I pointed out some things that were bad about AI. Basically none of them have changed. Many things have gotten worse. Among other things, the attack on labor has intensified. Here's why to resist dehumanizing your work: www.acsu.buffalo.edu/~cxjacobs/Bl...

1 week ago 38 15 2 1

folks please respect the sign’s privacy at this time

1 week ago 4 0 0 0
Post image

oh shit, oh sorry! oh god i’m making it worse now? Shit, sorry!

1 week ago 8 0 1 0

nice

2 weeks ago 2 0 0 0

in response to a survey asking authors if they intentionally put symbolism in their novels, Iris Murdoch said

“There is much more symbolism in ordinary life than some critics seem to realize.”

2 weeks ago 13 2 0 0

Microsoft, to its billions of customers

2 weeks ago 8 2 0 0
Advertisement

aww, thanks

2 weeks ago 2 0 1 0

fwiw, i started drafting my blog post about LLMs weeks ago and only got around to finishing it a few days ago. its timing with the AI guy crashing out is completely coincidental

2 weeks ago 7 0 0 0
Preview
LLMs are Not Just Another Research Automation – Væl Space

I’ve been thinking about the construal of LLMs as tools for research automation, and as someone who’s put a lot of effort into research facilitation tools, I don’t agree

2 weeks ago 27 6 2 1

Me and my sister’s Baldur’s Gate 3 campaign ended with her unleashing Artistry of War directly into Volo’s face at the after party, to which the game responded by cutting directly to the end credits

3 weeks ago 7 0 1 0

i’m curious about how AI cliches happen. Is “it’s not just X, it’s Y” just the median rhetorical device in rhetorical training data? does something about the reenforcement training, or the system prompt predispose this frame?

3 weeks ago 8 0 0 0

i don’t think being the governor of california licenses you to use the f slur like this

3 weeks ago 9 0 0 0

Wikipedia now has higher standards than all universities

3 weeks ago 5854 1766 30 40

the nosalgia of this

3 weeks ago 2 0 0 0

potential fixes

- belly beasties
- tummy buddies
- colorectal compadres
- the abdominal horde

4 weeks ago 2 0 1 0
Advertisement