fauxneme
@fauxnemic.bsky.social
language & mind-ish things (mostly fun sometimes sin), interested in what science is and does, pretend stoic but for-real epicurean (but for-real-real hedonist), we're in this together, la do la si, &c &c
tryna be drunk (comme Baudelaire)
tryna be drunk (comme Baudelaire)
i bought a(nother) bass
November 16, 2025 at 4:23 AM
i bought a(nother) bass
Reposted by fauxneme
Anyone interested in spending a semester to a year as a guest prof in Gothenburg? Let me know...
November 15, 2025 at 5:44 PM
Anyone interested in spending a semester to a year as a guest prof in Gothenburg? Let me know...
Reposted by fauxneme
Key point whenever you see an Altman prediction: the purpose of his predictions is *not* to accurately guess the future but build rationales in the present favorable to his company's interests. If you ignore everything he says you'll be smarter about AI.
Altman, and all these folks don't make predictions in order to predict the future, they make them to a) justify an overclocked corporate valuation, and b) argue for undoing regulations and receiving government support.
November 15, 2025 at 8:42 PM
Key point whenever you see an Altman prediction: the purpose of his predictions is *not* to accurately guess the future but build rationales in the present favorable to his company's interests. If you ignore everything he says you'll be smarter about AI.
who's got two thumbs and is on his way to see KRS-One
November 14, 2025 at 4:13 AM
who's got two thumbs and is on his way to see KRS-One
Reposted by fauxneme
“Register effects of imagined addressees on f0 across generations” by Miriam Oschkinat, Melanie Weirich, Daniel Duran and Stefanie Jannedy (Nov. ’25) pubs.aip.org/asa/jel/arti...
November 12, 2025 at 2:46 AM
“Register effects of imagined addressees on f0 across generations” by Miriam Oschkinat, Melanie Weirich, Daniel Duran and Stefanie Jannedy (Nov. ’25) pubs.aip.org/asa/jel/arti...
some account amusingly QTed this with
'not me immediately thinking "leonard bumstain"'
but they had replies and QTs disabled so now you have to merely imagine me QTing it unimaginatively with
"i think you meant bumstein"
'not me immediately thinking "leonard bumstain"'
but they had replies and QTs disabled so now you have to merely imagine me QTing it unimaginatively with
"i think you meant bumstein"
November 12, 2025 at 5:16 AM
some account amusingly QTed this with
'not me immediately thinking "leonard bumstain"'
but they had replies and QTs disabled so now you have to merely imagine me QTing it unimaginatively with
"i think you meant bumstein"
'not me immediately thinking "leonard bumstain"'
but they had replies and QTs disabled so now you have to merely imagine me QTing it unimaginatively with
"i think you meant bumstein"
i liked dynet (not theano)
For anyone that has worked in DL long enough to go from DIY to theano/chainer/tensorflow/mxnet/dynet/pytorch/Jax, one of these "winning" over all other felt unlikely ... and it kinda happened
November 10, 2025 at 4:12 AM
i liked dynet (not theano)
ok but like I'd rather eat a steak and drink a glass of milk than be encouraged to whack myself
Nebraska farmers alone use ~16 trillion liters a year. AI is a rounding error by comparison prospect.org/environment/...
Water Is Not the Problem With Artificial Intelligence
Data centers are nothing compared to cattle farming.
prospect.org
November 9, 2025 at 1:04 AM
ok but like I'd rather eat a steak and drink a glass of milk than be encouraged to whack myself
i took my first voice lesson today y'all; felt good
November 7, 2025 at 4:20 AM
i took my first voice lesson today y'all; felt good
I'm starting voice lessons next week
November 2, 2025 at 5:12 PM
I'm starting voice lessons next week
ok listen 7up Shirley Temple is absolutely the greatest thing to happen to mass-market soft drinks to date in the 21st century
October 31, 2025 at 8:24 PM
ok listen 7up Shirley Temple is absolutely the greatest thing to happen to mass-market soft drinks to date in the 21st century
did gmail's spam filters just shit the bed?
October 30, 2025 at 2:28 PM
did gmail's spam filters just shit the bed?
damn RIP Anthony Jackson
October 21, 2025 at 4:33 AM
damn RIP Anthony Jackson
Reposted by fauxneme
“Indeed, throughout modern society, the availability of statistics is often mistaken for the availability of knowledge and deep meaning.”
Todd Gitlin in 1979 with a line that feels profoundly prescient for our outsized contemporary worship of computer scientists and MBAs.
Todd Gitlin in 1979 with a line that feels profoundly prescient for our outsized contemporary worship of computer scientists and MBAs.
October 20, 2025 at 11:55 PM
“Indeed, throughout modern society, the availability of statistics is often mistaken for the availability of knowledge and deep meaning.”
Todd Gitlin in 1979 with a line that feels profoundly prescient for our outsized contemporary worship of computer scientists and MBAs.
Todd Gitlin in 1979 with a line that feels profoundly prescient for our outsized contemporary worship of computer scientists and MBAs.
Reposted by fauxneme
they should always let the pitchers hit tbh
October 18, 2025 at 2:02 AM
they should always let the pitchers hit tbh
Reposted by fauxneme
This can't be real can it? An AGI paper from many big names in the field that invents a metric of AGI and claims we're 58% of the way there, and then the references are fake, generated by an LLM? That's got to be an elaborate hoax doesn't it? It's too perfect.
The viral "Definition of AGI" paper tells you to read fake references which do not exist!
Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
Take this as a warning to not use LMs to generate your references!
Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
Take this as a warning to not use LMs to generate your references!
October 19, 2025 at 12:25 AM
This can't be real can it? An AGI paper from many big names in the field that invents a metric of AGI and claims we're 58% of the way there, and then the references are fake, generated by an LLM? That's got to be an elaborate hoax doesn't it? It's too perfect.
Reposted by fauxneme
On how counterfactual reasoning can break our normative complacency: "Historical observation and data-gathering are theory-laden activities as much as their scientific counterparts are. >>
October 18, 2025 at 3:32 AM
On how counterfactual reasoning can break our normative complacency: "Historical observation and data-gathering are theory-laden activities as much as their scientific counterparts are. >>
Reposted by fauxneme
"Although I hate leafy vegetables, I prefer daxes to blickets." Can you tell if daxes are leafy vegetables? LM's can't seem to! 📷
We investigate if LMs capture these inferences from connectives when they cannot rely on world knowledge.
New paper w/ Daniel, Will, @jessyjli.bsky.social
We investigate if LMs capture these inferences from connectives when they cannot rely on world knowledge.
New paper w/ Daniel, Will, @jessyjli.bsky.social
October 16, 2025 at 3:27 PM
"Although I hate leafy vegetables, I prefer daxes to blickets." Can you tell if daxes are leafy vegetables? LM's can't seem to! 📷
We investigate if LMs capture these inferences from connectives when they cannot rely on world knowledge.
New paper w/ Daniel, Will, @jessyjli.bsky.social
We investigate if LMs capture these inferences from connectives when they cannot rely on world knowledge.
New paper w/ Daniel, Will, @jessyjli.bsky.social
Reposted by fauxneme
Reposted by fauxneme
Sad. A tremendous linguist has left us: “The Yale Department of Linguistics is saddened at the news that our longtime colleague Stephen R. Anderson has died.” ling.yale.edu/posts/2025-1...
October 16, 2025 at 4:09 AM
Sad. A tremendous linguist has left us: “The Yale Department of Linguistics is saddened at the news that our longtime colleague Stephen R. Anderson has died.” ling.yale.edu/posts/2025-1...
Reposted by fauxneme
Tom Bombadil is a key character to understanding Lord of the Rings and the fact that most adaptations leave him out shows that most adapters don't really get the source material
Tell me your most unhinged literary opinion, as a little treat
October 14, 2025 at 1:24 AM
Tom Bombadil is a key character to understanding Lord of the Rings and the fact that most adaptations leave him out shows that most adapters don't really get the source material
Reposted by fauxneme
imagine doing bryan ferry dirty like this
Not proud of this but I thought U2 - With Or Without You and Roxy Music - More Than This were the same song
October 13, 2025 at 5:18 AM
imagine doing bryan ferry dirty like this
Reposted by fauxneme
Anyone who claims that Silicon Valley would be better with more humanities education has to grapple with the fact that Peter Thiel was a philosophy major.
October 12, 2025 at 4:28 PM
Anyone who claims that Silicon Valley would be better with more humanities education has to grapple with the fact that Peter Thiel was a philosophy major.
Reposted by fauxneme
Wikipedia wants you to know that John Maynard Keynes is not the lead vocalist of Tool. Just in case you were confused.
October 11, 2025 at 9:04 PM
Wikipedia wants you to know that John Maynard Keynes is not the lead vocalist of Tool. Just in case you were confused.