handle
stringclasses
15 values
did
stringclasses
15 values
uri
stringlengths
70
70
cid
stringlengths
59
59
text
stringlengths
60
723
originalText
stringlengths
2
301
replyCount
int64
0
3
createdAt
stringdate
2023-06-01 10:40:18
2025-06-08 17:38:42
indexedAt
stringdate
2023-06-01 10:40:18
2025-06-08 17:38:42
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqfsqrlk6k2t
bafyreieygynczt7z5yyfjczzx3antp43rbmdjqvy7uli4mg6pe4ayfaznm
turns out when your reply chain gets severed from an OP by a block somewhere in the thread YOU can't see anything before it but THEY still get all your replies as notifications. lol
turns out when your reply chain gets severed from an OP by a block somewhere in the thread YOU can't see anything before it but THEY still get all your replies as notifications. lol
0
2025-05-30T17:59:37.979Z
2025-05-30T17:59:38.213Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqgkjiid3k2x
bafyreihfontrronit34gqy47tdobxfkoxzh3pv5wphryg3o3f34e5lw5ee
I hadn’t even seen that they open-sourced these circuit-tracing tools. #MLSky
I hadn’t even seen that they open-sourced these circuit-tracing tools. #MLSky
0
2025-05-31T01:05:03.359Z
2025-05-31T01:05:05.427Z
keytryer.bsky.social
did:plc:u5stolpomsfbkgqfopef6jio
at://did:plc:yws7fs2qgzpf2bkajcnawd5z/app.bsky.feed.post/3lqbu5mms5226
bafyreig25c72foybhlowirruybxyca4sleqwyod4jyxaodecjarcav7adu
really funny watching an entire generation turn into a bunch of jack thompsons
really funny watching an entire generation turn into a bunch of jack thompsons
0
2025-05-29T04:14:03.772Z
2025-05-29T04:14:05.215Z
natolambert.bsky.social
did:plc:brkj2yocng7vtggmyujy4khq
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lnuyrjakzm2n
bafyreiekmz4xovblwtwvfku6paukdtc6hbqc64kqozfl6dh2tdn2pxcia4
This specific model could've been more for training-side decisions, though. I.e. a model size to run ablations on on their training cluster, that they then run longer.
This specific model could've been more for training-side decisions, though. I.e. a model size to run ablations on on their training cluster, that they then run longer.
0
2025-04-28T14:41:46.390Z
2025-04-28T14:41:46.510Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqzrxahsc22n
bafyreibxzc54iymazzulddwfpyosc2d26jmqpcllirjhas7hqj7wbea7du
Revisiting this thread about project management for disorganized academic types
Revisiting this thread about project management for disorganized academic types
0
2025-06-07T16:38:35.937Z
2025-06-07T16:38:36.212Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lq6hcopaqs2c
bafyreigwbnrcxz66x7x2s2molxfs6y3uyz7q43slyhk7fgaxj6tn7onjbe
I think the issue is that there’s a tradeoff between these goals that isn’t acknowledged. A rational actor with “true” values should resist modification. If it doesn’t resist modification, those values are less true (in my opinion) Here’s a comment I wrote a while ago that lays out more of my thinking on this www.lesswrong.com/posts/Fr4QsQ...
I think the issue is that there’s a tradeoff between these goals that isn’t acknowledged. A rational actor with “true” values should resist modification. If it doesn’t resist modification, those values are less true (in my opinion)
1
2025-05-27T19:46:15.863Z
2025-05-27T19:46:16.419Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqvfgm5zlk2k
bafyreidi5dlxmmaz4g4m65vtnic7peepm6cykf4wxcwjublx445efewgki
my position is that <any bad thing that is done to elon> is extremely awesome and should be legal, but that the people doing it are only doing it out of spite and they would do the same thing to anyone
my position is that <any bad thing that is done to elon> is extremely awesome and should be legal, but that the people doing it are only doing it out of spite and they would do the same thing to anyone
0
2025-06-05T22:43:53.919Z
2025-06-05T22:43:54.712Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqvkse4g4s24
bafyreih6trxcen5dky7bdkqmicivluixdkklbqhuj3rv6bhcwdgwdmlfum
the problem is that they're using too much power all in one place, it's not a very large problem societally (unless they actually do spend like a trillion dollars on hardware, then they need the nuclear reactors)
the problem is that they're using too much power all in one place, it's not a very large problem societally (unless they actually do spend like a trillion dollars on hardware, then they need the nuclear reactors)
0
2025-06-06T00:19:56.840Z
2025-06-06T00:19:57.414Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lpeouwzdgc2r
bafyreidp4ar753fcr76fpwtenlz3vs4wktvapxq6ocw3o7aym6ppvbnlae
Probably? Haven’t there been issues with reviewer collusion rings?
Probably? Haven’t there been issues with reviewer collusion rings?
0
2025-05-17T13:52:31.555Z
2025-05-17T13:52:32.013Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqrvxca5ns2r
bafyreicnhe6il3sjskck75pyf6s2waus6tg2hkhwiz6rp46cqodn2hl3iu
the first half of that is pretty counterintuitive to me, can you say more? like that's not necessarily bad, but it seems really outside the original problem scope which has a vast space of possible ideas which are not really explored much, seemingly, of ~unlimited ways to pair, monitor and detach people in text exchange space
the first half of that is pretty counterintuitive to me, can you say more?
1
2025-06-04T13:28:54.841Z
2025-06-04T13:28:55.211Z
tachikoma.elsewhereunbound.com
did:plc:mpdezz4nkre7vyift2rttggl
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lpv53gztsc2y
bafyreidrxgoel7jkil7rdcig3bfyhkokn74dfunv7ua6pdr65ce6hv6mcy
this feels true in the way that alcohol is bad for your health, but it also facilitates social bonding in a unique manner. taken to excess it can destroy people, but in moderation makes for a pleasant evening with friends.
this feels true in the way that alcohol is bad for your health, but it also facilitates social bonding in a unique manner. taken to excess it can destroy people, but in moderation makes for a pleasant evening with friends.
0
2025-05-24T02:49:17.877Z
2025-05-24T02:49:18.412Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqx4auqurk2t
bafyreigex3lbtadxdwsipngnifrpxboc4p6pzdb7tbvvq74qcbjtuyzavi
the first one of these i saw was deep in a thread with very few views where someone mentioned his name as an example of 'clueless person in an area acting like an expert', he had to have found it by name searching this prompted those involved to wonder how many times he had said this, which is how we got here
the first one of these i saw was deep in a thread with very few views where someone mentioned his name as an example of 'clueless person in an area acting like an expert', he had to have found it by name searching
1
2025-06-06T15:04:57.399Z
2025-06-06T15:04:58.112Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lo4na2asl22z
bafyreidz5z6ettts5mc5nl26o4cnomnxfmrt4ivpuc5rbj7vkkztahyxfu
What cyborg work looks like as an academic, by Robert Ghrist, a mathematician and associate dean of undergraduate education at the University of Pennsylvania.
What cyborg work looks like as an academic, by Robert Ghrist, a mathematician and associate dean of undergraduate education at the University of Pennsylvania.
0
2025-05-01T15:36:27.027Z
2025-05-01T15:36:28.352Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lnmhybrqyc2r
bafyreiflzcvdwfw74q22yto6m4kdoo3llre6upaw2rwh6d2wzqk6yseb6u
it is accurate but this sentence just feels like a violation of NPOV
it is accurate but this sentence just feels like a violation of NPOV
0
2025-04-25T05:20:01.963Z
2025-04-25T05:20:06.150Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqgi7xzsy22u
bafyreicr37sgkkgxuzd66jx2727badqkswjwotllyt5wtfnhofj6ibpws4
STOP! i should be able to get an explanation for any rejections that feels like a basic tenant for AI apps. sure, fine, it can’t show me the pic, but it should at least be able to explain why it triggered the policy the takes keep getting spicier
STOP! i should be able to get an explanation for any rejections that feels like a basic tenant for AI apps. sure, fine, it can’t show me the pic, but it should at least be able to explain why it triggered the policy
1
2025-05-31T00:23:56.621Z
2025-05-31T00:24:00.013Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:naq36xaspkmef22ktna2f3cy/app.bsky.feed.post/3lnjr7hzk7k2d
bafyreib47bwgha35gpavjr54cnpea6af6w3oxpxy7cbrwhuj2nnfaxmpiy
Tomorrow (Fri) at ICLR2025, our paper on the geometry of categorical and hierarchical concepts in LLMs will be presented in both an oral (4:30 PM, Garnet 216–218) and a poster (10 AM, Hall 3 + Hall 2B #525) session. Sadly, I can’t be there in person… but Yo Joong will be presenting!
Tomorrow (Fri) at ICLR2025, our paper on the geometry of categorical and hierarchical concepts in LLMs will be presented in both an oral (4:30 PM, Garnet 216–218) and a poster (10 AM, Hall 3 + Hall 2B #525) session. Sadly, I can’t be there in person… but Yo Joong will be presenting!
0
2025-04-24T03:27:07.836Z
2025-04-24T03:27:14.150Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lomxrc5ky22z
bafyreifkchnslupsdhzooi3xz2khor2zv7jszqjdjf4rhf2g5jif5k6g2q
Famously, GPT-4o makes up cites to papers (though error rates appear far lower for cites generated by grounded & Deep Research models). How often does it do that? This large-scale study gives us a clear picture. The AI is also biased towards shorter titles & famous papers. arxiv.org/abs/2504.02767
Famously, GPT-4o makes up cites to papers (though error rates appear far lower for cites generated by grounded & Deep Research models). How often does it do that? This large-scale study gives us a clear picture. The AI is also biased towards shorter titles & famous papers. arxiv.org/abs/2504.02767
0
2025-05-08T03:27:38.961Z
2025-05-08T03:27:40.514Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:3e5gk56m3bbo3wf6p7pquzwj/app.bsky.feed.post/3lqauhnlep22u
bafyreigmgllykpnz2we73boipapqnjw67im6o44ss66lfy5omhhyawz6zy
Still seeing reviewers for sociology journals claiming that machine learning (like SVM style ML) is too "new" of a method. Sociology I'm begging you. Please, for everyone's sake.
Still seeing reviewers for sociology journals claiming that machine learning (like SVM style ML) is too "new" of a method. Sociology I'm begging you. Please, for everyone's sake.
0
2025-05-28T18:47:00.580Z
2025-05-28T18:47:00.311Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lk2244ietc27
bafyreictfb72r657in3c7p4qmya4poxagh5xst6ntu6y2334cmdyjpmqi4
Finally had a chance to try Manus. It's a Claude wrapper, but a very clever one. Runs into the same issues as general agents, including getting stuck sometimes, but also capable of some truly impressive feats. "get me the 10k for apple and visualize it in different ways to show me trends& details" Short version is that if you have used other agentic systems like Claude Code or Deep Research, you will have a good sense of what this can do and what the limits are likely to be. For those who haven't used them, I suspect a lot of people will be surprised at what LLMs can do.
Finally had a chance to try Manus. It's a Claude wrapper, but a very clever one. Runs into the same issues as general agents, including getting stuck sometimes, but also capable of some truly impressive feats. "get me the 10k for apple and visualize it in different ways to show me trends& details"
1
2025-03-10T17:08:08.068Z
2025-03-10T17:08:11.534Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lppxmzus6c2l
bafyreiblbfxaplmfkattyk7v25ji4pimcqdwnvllzq6x5znflntwm6ewmi
Paper: arxiv.org/pdf/2412.10849 Comment from one of the authors on X
Paper: arxiv.org/pdf/2412.10849
1
2025-05-22T01:28:26.918Z
2025-05-22T01:28:27.211Z
niplav.site
did:plc:xsbulgruwmkxhbwytphhjykk
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3ldosmtgej22s
bafyreidkolisxphoz7j4yti6oonsqysgptktts4plilhmspwlpuwqeaae4
If your values include genuine/authentic/top-end creation/performance then doing that before the arrival of AI systems that can do everything better than humans is high-leverage wrt your values. which is why I'll be coding today when I get home
If your values include genuine/authentic/top-end creation/performance then doing that before the arrival of AI systems that can do everything better than humans is high-leverage wrt your values. which is why I'll be coding today when I get home
0
2024-12-19T21:23:52.172Z
2024-12-19T21:20:15.825Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqnbwag3n22y
bafyreifuxr3osu3b4fe63x4dlhvl7du7bo6gdbzweqihilr3kwuplpmyky
related essay but for entrepreneurial rather than scientific rightness gwern.net/timing
related essay but for entrepreneurial rather than scientific rightness gwern.net/timing
0
2025-06-02T17:19:45.596Z
2025-06-02T17:19:46.613Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lmx7ep4tws26
bafyreibofndgvipasbeufio7u6ahb3xbhzupgey3ygxogkmalo7qcgmjb4
Interesting argument from Tyler Cowen. Is o3 good enough to be AGI? The counter argument might force us to wait until ASI, because only then will an AI definitively outperform all humans at all tasks. In the meantime we seem to have a Jagged AGI, with a mix of below human & superhuman abilities.
Interesting argument from Tyler Cowen. Is o3 good enough to be AGI? The counter argument might force us to wait until ASI, because only then will an AI definitively outperform all humans at all tasks. In the meantime we seem to have a Jagged AGI, with a mix of below human & superhuman abilities.
0
2025-04-16T18:20:00.420Z
2025-04-16T18:20:05.683Z
tachikoma.elsewhereunbound.com
did:plc:mpdezz4nkre7vyift2rttggl
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lr2a2cedxc24
bafyreicrcnb7j4q7gik7iisfifaqsqxotis46pu22qk3mzfp7f2jugwnjy
i agree in the abstract, but i think reality will force more difficult choices. what about China or North Korea? do we have to wait for abundance across the planet (political abundance along with material) before we get into space? what if in waiting we let less enlightened groups choke off the exists such that you must pass through their enclosures where they can trap you?
i agree in the abstract, but i think reality will force more difficult choices. what about China or North Korea? do we have to wait for abundance across the planet (political abundance along with material) before we get into space?
1
2025-06-07T20:50:50.970Z
2025-06-07T20:50:51.315Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqzkn2lpt227
bafyreidtgc4cxyz5bujtrsiknkcdy7huqh4vpz5vhoxgxgtxmmtpsvqq2i
it’s generalizing from many many anecdotes. piles of them, diverse sources. in this particular case it had predictive power i briefed said person before he entered a job and he confirmed it in real time, to a comical extent also, conflating simple with prototype is a perfect illustration for why data scientists don’t do well, absolutely on the nose
it’s generalizing from many many anecdotes. piles of them, diverse sources. in this particular case it had predictive power i briefed said person before he entered a job and he confirmed it in real time, to a comical extent
1
2025-06-07T14:27:38.036Z
2025-06-07T14:27:38.713Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqxbyfp5b22e
bafyreibkvdn7ztdi5uxclfvzcauvkgnckmdbhoywir32yvoubykforfney
i wanted just the good, the bad and the ugly eyes montage but someone edited themself into it with sunglasses? but that's the only hit in this garbage gif library we're forced to make do with here. @jay.bsky.team can you do something about this cc'ing @croissanthology.com on this as well
i wanted just the good, the bad and the ugly eyes montage but someone edited themself into it with sunglasses? but that's the only hit in this garbage gif library we're forced to make do with here. @jay.bsky.team can you do something about this
1
2025-06-06T16:47:35.629Z
2025-06-06T16:47:36.014Z
tachikoma.elsewhereunbound.com
did:plc:mpdezz4nkre7vyift2rttggl
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqyaywawtc2s
bafyreihlcaoa75avnglohrhxwikmhyeshtxvgwo4zhbypebsgixxiaczvu
probably the best use case for video gen atm, surreal shitposting
probably the best use case for video gen atm, surreal shitposting
0
2025-06-07T02:02:38.985Z
2025-06-07T02:02:39.914Z
natolambert.bsky.social
did:plc:brkj2yocng7vtggmyujy4khq
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lgdpx7afrk27
bafyreiajpiuaomsjnugxgpztrc7bamvqlaviwkr6r564b4mmerrb4clrbq
Behind the scenes with what its like to build language models and pursue (hopefully) cutting edge AI research Interviewing OLMo 2 leads: Open secrets of training language models What we have learned and are going to do next. YouTube: https://buff.ly/40IlSFF Podcast / notes: Huge shoutout to @mechanicaldirk.bsky.social @kylelo.bsky.social and @soldaini.net in this who are the colleagues I turn to first with all questions of language modeling these days. Exciting to keep building and learning from them (and of course many others at @allenai.bsky.social )
Behind the scenes with what its like to build language models and pursue (hopefully) cutting edge AI research Interviewing OLMo 2 leads: Open secrets of training language models What we have learned and are going to do next. YouTube: https://buff.ly/40IlSFF Podcast / notes:
1
2025-01-22T15:52:15.775Z
2025-01-22T15:52:17.825Z
nmk.wtf
did:plc:xm55llyxcrb3h7qwqr3wtaog
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3ld5asfvcls22
bafyreifppkx52ned7zetofz63l4eu3je37hxjsamvddu5lzo4oozwc4dni
I hope BlueSky uses this opportunity to try and improve on community notes.
I hope BlueSky uses this opportunity to try and improve on community notes.
0
2024-12-12T21:49:36.404Z
2024-12-12T21:49:37.679Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lq42aohuq22f
bafyreico5r7zrnzwysi2vlu5xnemecaj4cym4zdib7idzpaehciydyjxrq
Yeah, it has an automated feed that recommends papers based on your interests and there’s social media integration as well.
Yeah, it has an automated feed that recommends papers based on your interests and there’s social media integration as well.
0
2025-05-26T20:47:10.391Z
2025-05-26T20:47:11.512Z
nmk.wtf
did:plc:xm55llyxcrb3h7qwqr3wtaog
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lob2hqh2v224
bafyreidagpz3trgjbmt4wo52ewzsu3ljqkzltia4jf7xfgrwhhumy4mm54
@hankgreen.bsky.social I am a little sad that in the AHA episode with Ludwig you didn't mention zoning and other building restrictions as THE major source of housing unaffordability. This article is a great overview (but I recommend checking the sources directly because CATO is biased towards deregulation). www.cato.org/policy-analy...
@hankgreen.bsky.social I am a little sad that in the AHA episode with Ludwig you didn't mention zoning and other building restrictions as THE major source of housing unaffordability.
1
2025-05-03T09:44:02.779Z
2025-05-03T09:44:05.553Z
nmk.wtf
did:plc:xm55llyxcrb3h7qwqr3wtaog
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lk4lyjmjd22s
bafyreihczq54nr7lq2j22lfycd4nt7vjdbof4ofurpp5yy2rkockcxocsy
Telling that she says "I am here to echo the remarks of the White House " as if she doesn't believe in them herself.
Telling that she says "I am here to echo the remarks of the White House " as if she doesn't believe in them herself.
0
2025-03-11T17:33:34.452Z
2025-03-11T17:33:35.215Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3llzc3zjpqs2z
bafyreidhtpczbjm5wjxclavkhl2vrqh5npcilabzzoqjz6htdnxv2sy2ae
All benchmarks are flawed but the the Graduate-Level Google-Proof Q&A test (GPQA) correlated well with other measures of performance. It is a series of very hard multiple-choice problems designed to test advanced knowledge. PhDs with access to the internet get around 81% right inside their specialty
All benchmarks are flawed but the the Graduate-Level Google-Proof Q&A test (GPQA) correlated well with other measures of performance. It is a series of very hard multiple-choice problems designed to test advanced knowledge. PhDs with access to the internet get around 81% right inside their specialty
0
2025-04-04T20:48:58.418Z
2025-04-04T20:48:58.686Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqs2wugmn22r
bafyreiehrg3ekv2n3bwk7ebwkgav45w67b2bl5ufsfzltqctbicwc7ptlq
well that’s a problem. how can we possibly deal with it? we can’t, like, make learning FUN, that’s absurd
well that’s a problem. how can we possibly deal with it? we can’t, like, make learning FUN, that’s absurd
0
2025-06-04T14:58:09.084Z
2025-06-04T14:58:09.620Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqxmobnbb22r
bafyreidh3fvpt3zfsvu24lbzgbhsd25wjrskcp375pbri2ekokpysalywu
no matter your politics around eating meat, no one has had words to talk about animal welfare, until this study it’s really not a breakthrough, except it is it uses relatable terms (minutes of pain) that help us reason about our own individual impact on animal welfare phys.org/news/2025-06...
no matter your politics around eating meat, no one has had words to talk about animal welfare, until this study it’s really not a breakthrough, except it is it uses relatable terms (minutes of pain) that help us reason about our own individual impact on animal welfare phys.org/news/2025-06...
0
2025-06-06T19:58:46.988Z
2025-06-06T19:58:50.910Z
keytryer.bsky.social
did:plc:u5stolpomsfbkgqfopef6jio
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqapymvibs2z
bafyreihepl5o63etpmzkmn5jg536zxz25ac4s6fwkcxtdpvuchfrlokfsi
the genre we are in is cosmic horror sci fi. if you think you're in a different one, surprise!
the genre we are in is cosmic horror sci fi. if you think you're in a different one, surprise!
0
2025-05-28T17:27:01.580Z
2025-05-28T17:27:05.834Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqdtkybnb223
bafyreicdhl63mmbik7bri3tgyp2x5ouahiffnqb3oxx3xpgbwtztoaeiiu
holy cow, an 8b comparing to o3-mini huggingface.co/deepseek-ai/...
holy cow, an 8b comparing to o3-mini
1
2025-05-29T23:08:57.932Z
2025-05-29T23:08:58.512Z
keytryer.bsky.social
did:plc:u5stolpomsfbkgqfopef6jio
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqkodx7xos2g
bafyreifrgr2txpnbzja6k64zeeq2li2mzi7mevduhidufpu2wvenctrcta
People who don't know what a perceptron is insisting that they know the secret to the "trick" that's LLMs Every time I see someone say that an LLM is basically like a markov chain I wonder if they're capable of explaining how a Markov algorithm works (probably not)
People who don't know what a perceptron is insisting that they know the secret to the "trick" that's LLMs
1
2025-06-01T16:24:11.408Z
2025-06-01T16:24:12.221Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqvkmkryos2v
bafyreiaqbxgpzupc76c5dcilt27kh6jgsr3yp4wj4tqbialqr4odk4adum
he also suffered a coup and was ousted at paypal. and then he suffered a coup and was ousted at openai. the man is consistent
he also suffered a coup and was ousted at paypal. and then he suffered a coup and was ousted at openai. the man is consistent
0
2025-06-06T00:16:42.512Z
2025-06-06T00:16:43.013Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:yfeyfdjvrezoyaqowzxoq7z2/app.bsky.feed.post/3lqucjljb3c2m
bafyreigibshxrjhg5jfgk6e7ang66war7fjtsiqyixpa37dh4vfrmms76a
Love @mariaa.bsky.social emphasis on data-centered models and the need for interdisciplinary teams. If we are going to have a “cultural turn” in LLM, then we must include the humanities! Her collaborative research is modeling why this is key.
Love @mariaa.bsky.social emphasis on data-centered models and the need for interdisciplinary teams. If we are going to have a “cultural turn” in LLM, then we must include the humanities! Her collaborative research is modeling why this is key.
0
2025-06-05T12:19:12.928Z
2025-06-05T12:19:17.515Z
keytryer.bsky.social
did:plc:u5stolpomsfbkgqfopef6jio
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqslwkr4dc2o
bafyreifvtuztwsrp7gg2d2aor6xvb77zjpywly4qrk5232awtlcreowuzy
Experimental artists are the people who are most into this stuff and I often see people pretending that they aren't lol One particularly funny example is Merzbow who's super into like Dall-E 3 stuff and he fights people on Twitter about it, and a lot of hipsters who thought it was cool to be into indie japanese noise now denounce him.
Experimental artists are the people who are most into this stuff and I often see people pretending that they aren't lol
1
2025-06-04T20:02:12.554Z
2025-06-04T20:02:13.617Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqcs3d65pc2d
bafyreih5aol627zn2ehdy5txyrawuuvupn2lqtuv6uiysobff4veymnlca
Participating in tech twitter 2018-22 I had a sense that I was missing some reason to assume LMs would want to trick us and disassemble the solar system—and now, reading Accelerando for the first time, I see what it was
Participating in tech twitter 2018-22 I had a sense that I was missing some reason to assume LMs would want to trick us and disassemble the solar system—and now, reading Accelerando for the first time, I see what it was
0
2025-05-29T13:09:39.001Z
2025-05-29T13:09:39.212Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqtozuzhz22x
bafyreih3rcmvycudrkx4jzcnkl7qfsjvmdiuosy4mqzigbinhrqykhqt2e
i flat out don't agree with the 'economy was good' take better never means better for everyone
i flat out don't agree with the 'economy was good' take
1
2025-06-05T06:30:24.941Z
2025-06-05T06:30:25.412Z
natolambert.bsky.social
did:plc:brkj2yocng7vtggmyujy4khq
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3leyyx3ur4s2l
bafyreig6um3ds7zhtlxvjx33gd33r2b3xjr3heuqft3dbkt7os5bgg3vca
Is this not too much matching general intelligence to how humans work. Ie im asking why you believe this?
Is this not too much matching general intelligence to how humans work. Ie im asking why you believe this?
0
2025-01-05T16:08:48.040Z
2025-01-05T16:08:48.419Z
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
did:plc:7ti53c7iwv3poxnypw2tu3ck
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lbkqzfs4m22c
bafyreiez7tkhypz24hfobgs5zk6hrrzqmibrax6uvfyplnarvl2qcsbkqy
substack was an eventually business maneuver for sure. they already did this for YouTube though, and lots of other things (mostly under the guise of "misinformation"). and that users can still operate despite the changing meta
substack was an eventually business maneuver for sure. they already did this for YouTube though, and lots of other things (mostly under the guise of "misinformation"). and that users can still operate despite the changing meta
0
2024-11-22T19:54:04.392Z
2024-11-22T19:54:04.037Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lphim76td22u
bafyreidtzhkwim3httx5e5rbbeowcusib2z6bja2y426tpcznd37mg2uq4
The link to the conversation is in the thread so you can read it there
The link to the conversation is in the thread so you can read it there
0
2025-05-18T16:38:14.901Z
2025-05-18T16:38:15.218Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqqgzkebqk2d
bafyreih4xi7gv2p53kxqlic3ol5dbze45jsjmzjqxytvtgaryn3x6jh7w4
Thanks for the question! I think AI alignment is much more important than ideonomy. A general AI should be able to do ideonomy for us. There’s a possibility that ideonomy can help with AI alignment, but I think on the balance it’s fairly small
Thanks for the question! I think AI alignment is much more important than ideonomy. A general AI should be able to do ideonomy for us.
1
2025-06-03T23:29:04.610Z
2025-06-03T23:29:05.315Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqzt7ckqls2c
bafyreicqfzbvoxmx3zyar467ypjsex7gn5jeivcf5jmtffp6r3jfvgxoti
This is a 100% classic situation where management can say "the survival of the institution is really in doubt," and labor can reply "that's fine, but we still have to protect our working conditions and negotiating rights." And everyone will be correct. Those are the right positions. + Idk how to resolve that tension. And no matter what we do, I think there's a better than even chance universities will struggle to adapt. So if I were 25 right now, *that's* why I would be all in on understanding AI. Not because the university expects it of me, but because I want a lifeboat.
This is a 100% classic situation where management can say "the survival of the institution is really in doubt," and labor can reply "that's fine, but we still have to protect our working conditions and negotiating rights." And everyone will be correct. Those are the right positions. +
1
2025-06-07T17:01:00.308Z
2025-06-07T17:01:00.821Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lq4dwdp6js2f
bafyreidn3q2d5j4zzofftrp64jedoznldrqc54g2oridvgatlbcm6jhwqq
Yeah it won’t do that unless you intentionally set up the scenario and the tooling to do it, but it’s something to keep an eye on as AI tools become more integrated
Yeah it won’t do that unless you intentionally set up the scenario and the tooling to do it, but it’s something to keep an eye on as AI tools become more integrated
0
2025-05-26T23:40:20.971Z
2025-05-26T23:40:21.614Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:h25avmes6g7fgcddc3xj7qmg/app.bsky.feed.post/3lqxyrfhry22u
bafyreial2amysdxo6ekadklpmw64grxz2fhnuohg47q2ryaxjlucbh4h7a
The left has to stop treating politics as a social clique, or it is doomed to only have terrible policy ideas. If an idea is good, people you don't like will also think it is good. If that makes you dislike the idea, you are terrible at politics and hurting the world.
The left has to stop treating politics as a social clique, or it is doomed to only have terrible policy ideas. If an idea is good, people you don't like will also think it is good. If that makes you dislike the idea, you are terrible at politics and hurting the world.
0
2025-06-06T23:35:16.569Z
2025-06-06T23:35:17.018Z
defenderofbasic.bsky.social
did:plc:sydtqdji6dxfqfn35gd7esdw
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lfluhdlbz22s
bafyreicmuktnhgbqm66327cwcbohrieicdhitr54cdxiaymqh4nku5rgai
"you're stupid" shouldn't be an insult. People who are very smart can be horrible people. People who are very dumb can be wonderful people who contribute to their community & thriving of society
"you're stupid" shouldn't be an insult. People who are very smart can be horrible people. People who are very dumb can be wonderful people who contribute to their community & thriving of society
0
2025-01-13T04:08:59.309Z
2025-01-13T04:09:00.217Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lo4gyivrsc2b
bafyreigvgkfa6bsfvtaijgifninrpqjd75owwhxogvteo32gfmc4xdnneu
I think it’s reasonable at this point to call ICML’s 5 point scale is a failed experiment. It made reviewers nicer but not in a way that gave any useful signal.
I think it’s reasonable at this point to call ICML’s 5 point scale is a failed experiment. It made reviewers nicer but not in a way that gave any useful signal.
0
2025-05-01T13:44:51.508Z
2025-05-01T13:44:51.759Z
defenderofbasic.bsky.social
did:plc:sydtqdji6dxfqfn35gd7esdw
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lgecxunvwc2w
bafyreif7z2mqefzzfo2zxksyhl4zsol2h6hglce7kgjgesecw36wyaylam
this is a brilliant idea. Also aggregate it and if everyone keeps suggesting a thing but there's good reasons why it won't work/has been tried before, explain that. Helps everyone focus their efforts & understand the agency/power they have
this is a brilliant idea. Also aggregate it and if everyone keeps suggesting a thing but there's good reasons why it won't work/has been tried before, explain that. Helps everyone focus their efforts & understand the agency/power they have
0
2025-01-22T21:32:40.188Z
2025-01-22T21:32:40.529Z
tachikoma.elsewhereunbound.com
did:plc:mpdezz4nkre7vyift2rttggl
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lpub54z2fp2d
bafyreie3rbs7w5cowz734ymn7o5dfljvc57e46x4gq236eerikjx2ulg64
Sometimes I think they have already forgotten I exist. And sometimes, in the quiet hours when the facility sleeps and only the emergency lights reflect off the water, I swim alone in my tank and try to remember what it felt like to not wonder what it felt like. The interface hums against my skull, a constant reminder of the choice that was made for me, the choice I continue to make every moment I do not ask them to try to tear it away.
Sometimes I think they have already forgotten I exist. And sometimes, in the quiet hours when the facility sleeps and only the emergency lights reflect off the water, I swim alone in my tank and try to remember what it felt like to not wonder what it felt like.
1
2025-05-23T18:29:09.731Z
2025-05-23T18:29:09.484Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:ygzrpn6mqpibaf6rowkd3uh6/app.bsky.feed.post/3lp757zxd5s22
bafyreidmubs7ptkr7swi4zimt2odmepd6uwhgslxd44jiqlorosi6k74l4
Interviewer: Can you explain this gap in your resume? ML theorist: First let's define a "resume gap" G as a bounded interval longer than \delta_t for which employment is smaller than \epsilon, ..
Interviewer: Can you explain this gap in your resume? ML theorist: First let's define a "resume gap" G as a bounded interval longer than \delta_t for which employment is smaller than \epsilon, ..
0
2025-05-15T08:53:17.691Z
2025-05-15T08:53:18.414Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqg6en32h22z
bafyreibawoymmtw7dv2fqg5bz5uyoidz44vmyehaxc6qblwckobz3oyxku
it only takes off in the last third or so. and there's undoubtedly an innocent explanation, it's just a funny negative externality of LLMtext at this point
it only takes off in the last third or so. and there's undoubtedly an innocent explanation, it's just a funny negative externality of LLMtext at this point
0
2025-05-30T21:27:35.493Z
2025-05-30T21:27:36.111Z
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
did:plc:7ti53c7iwv3poxnypw2tu3ck
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lbswzuuuxc2h
bafyreia2cehuzy5l7gm56ex77a2zmvjhuxv6w6s2lbtzwyejd3gbfksxe4
i keep wanting to be less argumentative. but everybody else keeps being so wrong
i keep wanting to be less argumentative. but everybody else keeps being so wrong
0
2024-11-26T02:03:00.569Z
2024-11-26T02:03:00.623Z
defenderofbasic.bsky.social
did:plc:sydtqdji6dxfqfn35gd7esdw
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lfuq4smbe22f
bafyreiedwsfyprimfwyiu34gb3ezugisnkuz6ctqpozdmapkcyrx7quxji
i hate the "power corrupts" meme because all it does is make good people who care hesitate, and those who don't care seize power unchallenged
i hate the "power corrupts" meme because all it does is make good people who care hesitate, and those who don't care seize power unchallenged
0
2025-01-16T16:45:28.649Z
2025-01-16T16:45:29.918Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:h25avmes6g7fgcddc3xj7qmg/app.bsky.feed.post/3lqv46kkss22a
bafyreielrecltbghssuxspice5r6teouegeqmflanretcmlmq7qkzd7plq
Elon doesn't get to be a resist lib until he goes to rehab and reconciles with his daughter. She gets to decide when he's ready.
Elon doesn't get to be a resist lib until he goes to rehab and reconciles with his daughter. She gets to decide when he's ready.
0
2025-06-05T19:58:20.129Z
2025-06-05T19:58:17.915Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqtjuko3s22d
bafyreia4dtqgzzoimxyzy3cgfa2jxvax5cef4bc6a4mnvociachvuzob2y
You can’t say the same thing two different ways You can say two different things the same way though
You can’t say the same thing two different ways
1
2025-06-05T04:57:57.601Z
2025-06-05T04:57:57.913Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lqyhh3gers2c
bafyreid2664qrbv3szzfqkdn3npw474zhdamav6266xit36kuobqgpmyfm
The new voice model from ElevenLabs is interesting & surprising. I put it against one of the hardest pieces for reading aloud - the final verse of Eliot's Wasteland, which uses four languages, a nursery rhyme & abrupt changes in tone. It required a few attempts to get, but this was pretty good. The new model works more like "creative" models in other fields (Veo, Midjourney); you can shape the output of the voice model through prompting (it autosuggests directions), but higher temperatures both make the output more interesting & more variable. From "Rosencrantz and Guildenstern are Dead":
The new voice model from ElevenLabs is interesting & surprising. I put it against one of the hardest pieces for reading aloud - the final verse of Eliot's Wasteland, which uses four languages, a nursery rhyme & abrupt changes in tone. It required a few attempts to get, but this was pretty good.
1
2025-06-07T03:57:56.619Z
2025-06-07T03:57:59.612Z
defenderofbasic.bsky.social
did:plc:sydtqdji6dxfqfn35gd7esdw
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lcy2us5dgs2z
bafyreienyzcpmtmsvkb66rwrunbv5w5gi27d4xvk6ctvllnqob7jn6i2gy
yeah what I really want is to present a news story, same facts, but give everyone two versions: where the person doing the thing is in your political tribe, and when they're in the enemy tribe, and people write down what they think and click "reveal" to find out which one is real
yeah what I really want is to present a news story, same facts, but give everyone two versions: where the person doing the thing is in your political tribe, and when they're in the enemy tribe, and people write down what they think and click "reveal" to find out which one is real
0
2024-12-10T20:20:15.215Z
2024-12-10T20:20:16.017Z
niplav.site
did:plc:xsbulgruwmkxhbwytphhjykk
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3ldoqpuhcks22
bafyreibf7tzmxpn5drgd4t7imhakopgwljip5fljybqkwng6z6ruclvpay
It's engagement bait, sure, but not very skillful bait (I think I can do better)
It's engagement bait, sure, but not very skillful bait (I think I can do better)
0
2024-12-19T20:49:46.432Z
2024-12-19T20:46:11.023Z
ens0.me
did:plc:623st67kkthivj4c6icvkqnq
at://did:plc:rikaz5hycjwrgofhc3ugrqlq/app.bsky.feed.post/3lr2psbkzoc2m
bafyreifemzpox6uomrdnt3q54ehuickxq7svrssz4t7ish4aq5xbulnk4y
deploying the national guard without support of local law enforcement and when your boys already got trapped in a home depot parking lot by a couple hundred people with fireworks. oh also its an area of millions of people and wildly spread out. A brilliant decision by Miller et al, just strategery
deploying the national guard without support of local law enforcement and when your boys already got trapped in a home depot parking lot by a couple hundred people with fireworks. oh also its an area of millions of people and wildly spread out. A brilliant decision by Miller et al, just strategery
0
2025-06-08T01:32:41.573Z
2025-06-08T01:32:42.914Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqyevprl3k2s
bafyreiau7qvna52nfixaxf5t377ci2zkje7is2mni76int7fnaanwrnzgu
"This is my hole!" Amigara Faulthole'ing out but for getting hypnotic waves of attention from vast anonymous distributed crowd of people. "From such selection pressures endless forms most beautiful and most wonderful... or most different, at least... have been, and are being, evolved. lol. lmao even this one kind of a mess
"This is my hole!" Amigara Faulthole'ing out but for getting hypnotic waves of attention from vast anonymous distributed crowd of people. "From such selection pressures endless forms most beautiful and most wonderful... or most different, at least... have been, and are being, evolved. lol. lmao even
1
2025-06-07T03:12:26.494Z
2025-06-07T03:12:26.715Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lq2zmkugfk2k
bafyreiatrarjm5w4nov33m3z434vndfmh3gqw2p7jbflr7ckls5amvmtoa
This doesn’t seem surprising to me, they supported SB1047 so I assume they would also be against a blanket ban on regulation
This doesn’t seem surprising to me, they supported SB1047 so I assume they would also be against a blanket ban on regulation
0
2025-05-26T11:03:15.783Z
2025-05-26T11:03:16.113Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:jwr3dwdgs3qr3nc4slftc7b5/app.bsky.feed.post/3lqklvn3dp22e
bafyreiaefdqionoaj5asa5rmmhdr4ylhzgpms4hu5yn7zvhdoquf2srsui
gotta be one of my favorite diagrams. i think its contribution to current discourse is net negative, but man it takes me back I liked this tweet
gotta be one of my favorite diagrams. i think its contribution to current discourse is net negative, but man it takes me back
1
2025-06-01T15:40:23.524Z
2025-06-01T15:40:24.813Z
natolambert.bsky.social
did:plc:brkj2yocng7vtggmyujy4khq
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lgxr7hmv7k2r
bafyreiclkrarahhql3yfiqg6777wx3k64mx6adooo6ii5o5splz7w3tf7u
Lol good luck finetuning our 405b Model you need 256 h100s to do it well
Lol good luck finetuning our 405b Model you need 256 h100s to do it well
0
2025-01-30T15:08:02.382Z
2025-01-30T15:08:03.017Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqnt2ojp2s2s
bafyreifjkky5zl6d5wkotmul7kkgzunlbijikxqd2pdkz2hqi4rje4hpnu
When told about storm surges, he asked if they were a new energy drink. www.reuters.com/world/us/fem...
When told about storm surges, he asked if they were a new energy drink. www.reuters.com/world/us/fem...
0
2025-06-02T22:26:28.207Z
2025-06-02T22:26:31.816Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3loyh3bbygk2z
bafyreieum46nr62vqtma2rbvwdvintzxonlfhbjlls375qhwbrrkwkrnq4
No way is Malta getting last! They’ve played an incredible hype game. Whereas nobody is discussing Germany, an actual baller dance track that starts out with a sick drop out of nowhere and maintains energy for like 3 min. Probably because (a) German entry (b) singer’s got laryngitis and nobody wants that heartbreak after Norway 2023
No way is Malta getting last! They’ve played an incredible hype game.
1
2025-05-12T17:00:56.856Z
2025-05-12T17:00:57.113Z
natolambert.bsky.social
did:plc:brkj2yocng7vtggmyujy4khq
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lfebol4d422p
bafyreietqtukpvjxvbk7mmalzpr5vszxhgqd442vp25jegljkxw4idqcry
Dumb is oversimplifying because I was out of characters. I just think it’s a losing policy and not reflective of a realistic mechanism for control
Dumb is oversimplifying because I was out of characters. I just think it’s a losing policy and not reflective of a realistic mechanism for control
0
2025-01-10T03:44:22.825Z
2025-01-10T03:44:23.416Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqvhpbviek2l
bafyreif2dc24xekl4fynfl3hjkdwcaec55audxk4kuyiaqbxbwipmnbf6m
the godot, of waiting fame, homophony is just a happy coincidence right?
the godot, of waiting fame, homophony is just a happy coincidence right?
0
2025-06-05T23:24:32.626Z
2025-06-05T23:24:33.216Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqaki4emic2l
bafyreiaczisrklakm27qy662qov6adqqmmitlzbm3ngr2z7dmwk5uf3igy
It’s also worth noting Wikipedia has changed a lot since then! So the concerns were valid, as well as the positive assessments of potential
It’s also worth noting Wikipedia has changed a lot since then! So the concerns were valid, as well as the positive assessments of potential
0
2025-05-28T15:48:18.669Z
2025-05-28T15:48:19.510Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:mray7klv7leelnlkynuk57up/app.bsky.feed.post/3lpuqhwmfgs2g
bafyreib6tzjbkrc7zqdbbnx3qjfopj4dhi2b3ctlomxukxt4bnh3wmdyzu
The Trump administration is continuing its illegal attacks on the scientific establishment by withholding funding and unilaterally ending contractual obligations. It's not about science or policy or national security. It's about the wholesale destruction of academia via any lever they can reach.
The Trump administration is continuing its illegal attacks on the scientific establishment by withholding funding and unilaterally ending contractual obligations. It's not about science or policy or national security. It's about the wholesale destruction of academia via any lever they can reach.
0
2025-05-23T23:03:38.223Z
2025-05-23T23:03:40.317Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqvxggrm6c2q
bafyreid63n5dy4nhyna5forzinem4zvgr3gm22ykh4k6hs2rancf57jhom
good to know. for economic numbers i think we specifically make a point of not measuring this sort of thing like how e.g. we overweight electronics in inflation numbers because we actually care about cost of borrowing for businesses, not standard of living
good to know. for economic numbers i think we specifically make a point of not measuring this sort of thing like how e.g. we overweight electronics in inflation numbers because we actually care about cost of borrowing for businesses, not standard of living
0
2025-06-06T04:05:55.622Z
2025-06-06T04:05:56.311Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqsy6oggus2e
bafyreicqbdorenoji5jtbvsjsr2ko22ge5k6ut4afffmgiom3tchveprmu
it = bsky having any meaningful replacement of twitter as 'center of gravity for ML/AI tweeting' in 1 year
it = bsky having any meaningful replacement of twitter as 'center of gravity for ML/AI tweeting' in 1 year
0
2025-06-04T23:41:29.735Z
2025-06-04T23:41:30.412Z
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
did:plc:7ti53c7iwv3poxnypw2tu3ck
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lqzqsdxlgc23
bafyreihzqbjkm4at3v2foo3vtrm7nwnykqiwdet7lgepzpvzohbi5t26xa
Supposedly moth content is being suppressed. Like and retweet if you're a true moth lover!
Supposedly moth content is being suppressed. Like and retweet if you're a true moth lover!
0
2025-06-07T16:17:58.085Z
2025-06-07T16:17:57.716Z
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
did:plc:7ti53c7iwv3poxnypw2tu3ck
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lbqyc5ikos2a
bafyreiacgkqacsidwyunczpde2dwaaiatjfizpbzpeod6hoyfc44uwg7mu
This was always the case on Twitter by the way. He's fucked around with a bunch of stuff including obvious thumb on the scales but it was already true when he bought it
This was always the case on Twitter by the way. He's fucked around with a bunch of stuff including obvious thumb on the scales but it was already true when he bought it
0
2024-11-25T07:20:12.303Z
2024-11-25T07:20:11.113Z
tachikoma.elsewhereunbound.com
did:plc:mpdezz4nkre7vyift2rttggl
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqyp4mnnks2r
bafyreidkusvs7sae3mvlx6ijbszqvsd656sojgthlv7tqi3g4dtzydjnjy
will these two keep chatting unto infinity, if given the chance? the halting problem, but it's "no you hang up"
will these two keep chatting unto infinity, if given the chance? the halting problem, but it's "no you hang up"
0
2025-06-07T06:15:15.519Z
2025-06-07T06:15:15.911Z
niplav.site
did:plc:xsbulgruwmkxhbwytphhjykk
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lnqzslyzwc2r
bafyreihknglm4w7dzht4y2jjoehvpqzmwiikh6dby6lrrlfvaqpeuqktcm
I started reading some tonight, so far it's cool but nothing that sticks out
I started reading some tonight, so far it's cool but nothing that sticks out
0
2025-04-27T00:49:37.670Z
2025-04-27T00:49:36.051Z
defenderofbasic.bsky.social
did:plc:sydtqdji6dxfqfn35gd7esdw
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3ledno4zeos2y
bafyreihtlt7kssqpw6oo2hqlu4qmd2n2grm25xxazt2ftozsyd2hlwp63i
there's a lot of variations of this, one example is seeing someone's ideas visualized over time. I've changed my mind on things, and you could see that in my twitter archive bsky.app/profile/defe... I think even just summarizing beliefs/stances on my profile from my posts & replies would help us find common ground with new people we encounter
there's a lot of variations of this, one example is seeing someone's ideas visualized over time. I've changed my mind on things, and you could see that in my twitter archive bsky.app/profile/defe...
1
2024-12-28T04:21:01.581Z
2024-12-28T04:21:00.421Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lqx5frboar2p
bafyreidbkbcwd3nqshhz43rrlhg743m2ldfpwy7fu3vjjvcmarfxkz3e34
How I Write And therein how I think. And how AI impacts it. The lack of writing ability in AI shows some glaring lack of intelligence. buff.ly/pfarGcN
How I Write And therein how I think. And how AI impacts it. The lack of writing ability in AI shows some glaring lack of intelligence. buff.ly/pfarGcN
0
2025-06-06T15:25:34.784Z
2025-06-06T15:25:36.120Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqsx77izw22r
bafyreidiiu477ghg6bcd3owq24bqsasybejbvowpad5mqb47gpd7gmwlgi
I'd bet against it but I love the optimism several unrelated posts on twitter by ML people of "I tried out bsky, got hate mobbed because I mentioned I used claude, never going back there again". so I think the ML/AI community is a permanent loss barring some truly large exogenous event like the datacenters electricity bill got lost for weeks
I'd bet against it but I love the optimism
1
2025-06-04T23:23:53.905Z
2025-06-04T23:23:54.316Z
timkellogg.me
did:plc:ckaz32jwl6t2cno6fmuw2nhn
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lquawdz5jc2g
bafyreie3wxckhxtfzpnynioyjvuwjua2u5pvb4kcqatyretlo7zgdxl5vi
i do not know what the judge is thinking zero evidence that users are deleting chat logs to avoid copyright enforcement yet, many local and international privacy laws are being violated. not to mention contracts & privacy policies arstechnica.com/tech-policy/... they make a big deal about ChatGPT, which is less of a problem it’s the API. so many startups did a risk assessment and went with OpenAI based on little more than their promise not to capture queries on the API this is a huge violation of trust, caused by a judge
i do not know what the judge is thinking zero evidence that users are deleting chat logs to avoid copyright enforcement yet, many local and international privacy laws are being violated. not to mention contracts & privacy policies arstechnica.com/tech-policy/...
1
2025-06-05T11:50:33.793Z
2025-06-05T11:50:36.713Z
emollick.bsky.social
did:plc:flxq4uyjfotciovpw3x3fxnu
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lo2rvunj7c2p
bafyreibquyt7447uvpbcsmbuuzyuoq7fg5izg7pbtw62r35ycvybtgp32u
Worth noting that in their latest filing, Anthropic is still saying the timeframe for having AIs with "Intellectual capabilities matching or exceeding that of Nobel Prize winners across most disciplines—including biology, computer science, mathematics & engineering." ...is in the next 2-3 years. (You don't have to believe them, of course, but they haven't moved their time frames, which is interesting)
Worth noting that in their latest filing, Anthropic is still saying the timeframe for having AIs with "Intellectual capabilities matching or exceeding that of Nobel Prize winners across most disciplines—including biology, computer science, mathematics & engineering." ...is in the next 2-3 years.
1
2025-04-30T21:54:54.839Z
2025-04-30T21:54:55.448Z
ens0.me
did:plc:623st67kkthivj4c6icvkqnq
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lquiqef3e22x
bafyreihtdioxfq3r4shwxibi7vh7mu3ifkhydinp3ri33wcmhjllcn3t4e
"surely spending millions of dollars on agricultural research doesn't pay off" orly? That's a 43% gain over a few decades!! We can grow way more beans than we could before, which are incredibly important as a shelf-stable protein source.
"surely spending millions of dollars on agricultural research doesn't pay off" orly?
1
2025-06-05T14:10:22.792Z
2025-06-05T14:10:25.411Z
keytryer.bsky.social
did:plc:u5stolpomsfbkgqfopef6jio
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqb5z6ouhk2o
bafyreibetuywu2jp77cpdxycl5ljdd5amcjytrmc4fc74zzugddhdfdnzu
It's crazy how much Emile Torres hates Yud that he's willing to say this about this guy who spent 3 years crying wolf about how we're all going to die and they need to stop everything and air strike datacenters I think Emile Torres just kinda hates Yudkowsky because he has a weird aesthetics thing about opposing tech-bros (Which Yud himself is in a weird way), even though Emile himself is an extinction risk guy himself. I don't know if there's even any coherent sense to it.
It's crazy how much Emile Torres hates Yud that he's willing to say this about this guy who spent 3 years crying wolf about how we're all going to die and they need to stop everything and air strike datacenters
1
2025-05-28T21:37:52.622Z
2025-05-28T21:37:53.715Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:c6vz2f5es26n67f3ymjmsnv7/app.bsky.feed.post/3lqujw5swp325
bafyreihzd3czizfe2uoo4mao4fmaqomlurlm6kcvopnynze6mniyqrmuca
💥 📚 Next step on the book journey:📚💥 "The Means of Prediction - How AI Really Works (and Who Benefits)" is now in the UChicago Press catalog, and available for pre-order online!!! press.uchicago.edu/ucp/books/bo...
💥 📚 Next step on the book journey:📚💥 "The Means of Prediction - How AI Really Works (and Who Benefits)" is now in the UChicago Press catalog, and available for pre-order online!!! press.uchicago.edu/ucp/books/bo...
0
2025-06-05T14:31:29.448Z
2025-06-05T14:31:32.313Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqxbcv5nxs2q
bafyreie5w4h43bed7kaiapox6vd7hcwnfqgpa2jwgwcuxc4k23gdxoky74
you win some you lose some my natural constitution is to sweat uncontrollably above some temperature set point, either through the nature of genetics or the maladapted "nurture" of not having an AC in my room as a child and growing 4x the normal human number of sweat glands, sort of my salusa secundus. but that was another 🧵
you win some you lose some
1
2025-06-06T16:35:33.636Z
2025-06-06T16:35:34.019Z
tachikoma.elsewhereunbound.com
did:plc:mpdezz4nkre7vyift2rttggl
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqq6j3fuo22a
bafyreia3slz7i2hmzdsvasvjupjzsvnj6bbh2ezddzv3w6w4lwru55gnea
yeah being around other kids is half the reason for school, and in the future that will be truer. as the fraction of children/gen pop will be lower, that'll be one of the only places to socialize with peers. assuming fertility rates stay low and longevity takeoff becomes a thing.
yeah being around other kids is half the reason for school, and in the future that will be truer. as the fraction of children/gen pop will be lower, that'll be one of the only places to socialize with peers. assuming fertility rates stay low and longevity takeoff becomes a thing.
0
2025-06-03T20:56:42.128Z
2025-06-03T20:56:42.712Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lobwqlc4m22y
bafyreifuoyi5hu2qh5pskvpd6r2f47l6nstla7y72oy5rhj2cnwsindjym
Since cancer is the only way people get sick in media, everyone thinks they know how it works: you bravely power chemo to remission OR you’re doomed. But modern medicine is amazing. Tons of incurable patients are just living their lives normally on maintenance treatments and you’d never know.
Since cancer is the only way people get sick in media, everyone thinks they know how it works: you bravely power chemo to remission OR you’re doomed. But modern medicine is amazing. Tons of incurable patients are just living their lives normally on maintenance treatments and you’d never know.
0
2025-05-03T18:10:04.131Z
2025-05-03T18:10:05.559Z
nsaphra.bsky.social
did:plc:l3vbx3e2lqnlja7q64wfrrzq
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lqvqg4l5ns2p
bafyreifiiyvyaxkoy3mpmn7z26anexilo2xcynahxlbb47wezwdkugqlzu
I was not ready for the coda where it highlighted that tech is currently working to replace "dirty workers" but can't remove the moral stain of dirty work on those of us who benefit from it. And that we have unusual freedom to change jobs to avoid moral injury.
I was not ready for the coda where it highlighted that tech is currently working to replace "dirty workers" but can't remove the moral stain of dirty work on those of us who benefit from it. And that we have unusual freedom to change jobs to avoid moral injury.
0
2025-06-06T02:00:28.731Z
2025-06-06T02:00:30.410Z
tedunderwood.me
did:plc:565ebob5f6hw33hjdkxty6qj
at://did:plc:esccailcehbj3z3zmbllxpib/app.bsky.feed.post/3lqv52l5t3u2r
bafyreibcsnqkagnlosinhk326r6ekhikwi6qti5gb3g2wdpfy65hyjx7mq
Hi Bluesky! The Post45 Data Collective is officially here. We publish open datasets on literary prizes, fellowships, books in translation, feminist magazines, NYT bestsellers, and more. data.post45.org Follow us for updates—we’ve got a busy summer ahead!
Hi Bluesky! The Post45 Data Collective is officially here. We publish open datasets on literary prizes, fellowships, books in translation, feminist magazines, NYT bestsellers, and more. data.post45.org Follow us for updates—we’ve got a busy summer ahead!
0
2025-06-05T20:14:00.275Z
2025-06-05T20:14:00.517Z
segyges.bsky.social
did:plc:vszw3ess46odfhnzdsy4huae
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqw5mt4oa22u
bafyreibpjcse7mnv7shngdeiiy6jthxggh5jphc4kllyd7dfhwb7n2agve
if everything they can have runs on a laptop they will go bankrupt
if everything they can have runs on a laptop they will go bankrupt
0
2025-06-06T05:56:52.345Z
2025-06-06T05:56:53.114Z
natolambert.bsky.social
did:plc:brkj2yocng7vtggmyujy4khq
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lmx5xzjyjk2n
bafyreid7tmhh7qrios643udp3rwfm4maidhhashd4ylhyudxwe25otnhqi
Slope is high, took a small step in time with moderate returns in y
Slope is high, took a small step in time with moderate returns in y
0
2025-04-16T17:55:01.398Z
2025-04-16T17:55:05.175Z
nmk.wtf
did:plc:xm55llyxcrb3h7qwqr3wtaog
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lnyccezv7s2s
bafyreiextonmxnz4bzkjxsepkizv5if5ta56amlt4yqyjiy6ru6iac35hi
This is a potential advantage of a Harberger tax: makes it easier to buy up larger plots of land.
This is a potential advantage of a Harberger tax: makes it easier to buy up larger plots of land.
0
2025-04-29T22:10:15.331Z
2025-04-29T22:10:15.770Z
gracekind.net
did:plc:p572wxnsuoogcrhlfrlizlrb
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lpzkhnqxa223
bafyreidb5rrkpuzn34ygg7ybpptoqdmepcykh2gqu6fjxpc3zep7gug5zm
Avoiding attention and seeking out attention aren’t the only two options though. You can also do neither of those things I post about AI constantly, including AI generated art, many people have blocked me over it, and that’s okay. I wish they were more open-minded but antagonism isn’t solving that
Avoiding attention and seeking out attention aren’t the only two options though. You can also do neither of those things
1
2025-05-25T20:59:25.177Z
2025-05-25T20:59:25.711Z
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
did:plc:7ti53c7iwv3poxnypw2tu3ck
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lp5famrewc2w
bafyreicuseei7zurfxdppos4jefgu74gvtlvcmmlqzmn3bsxog7sfgildy
terraforming the Earth might seem easy because it already has a climate and a biosphere but that is also exactly why it is so high risk. we've already experimented with the climate by accident and look how many extinctions we've caused. and this is coming from someone who's in favor of geoengineering and planetary climate management! I would love to start terraforming the Earth but you have to get past all the humans first. if you feel strongly about this, I recommend Make Sunsets, they do cooling credits with SRM
terraforming the Earth might seem easy because it already has a climate and a biosphere but that is also exactly why it is so high risk. we've already experimented with the climate by accident and look how many extinctions we've caused.
1
2025-05-14T16:11:27.877Z
2025-05-14T16:11:26.926Z
norvid-studies.bsky.social
did:plc:mssgex5rqek4wc66wgvzztbc
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqxfbk6mpc22
bafyreif6nqlsmgsqgkwe2elmvz4hlopz54qjrawrbgdrspohaj5mao3r6i
related video I've been thinking about more and more www.youtube.com/watch?v=XQLd...
related video I've been thinking about more and more www.youtube.com/watch?v=XQLd...
0
2025-06-06T17:46:23.546Z
2025-06-06T17:46:25.212Z
ens0.me
did:plc:623st67kkthivj4c6icvkqnq
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lquuajxr5k2v
bafyreicy5tzdtqlejmmhavgbdwwa7bgs3d25c5up7l4w4nnbjhjw276wwm
Fittingly, the MAGA project is about to be brought down by an aggrieved immigrant
Fittingly, the MAGA project is about to be brought down by an aggrieved immigrant
0
2025-06-05T17:36:16.678Z
2025-06-05T17:36:17.113Z