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
|
|---|---|---|---|---|---|---|---|---|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqcpl4iyxk24
|
bafyreic7lmgakdfyyjlilsddurbfdvincp4k3kkgfryod76njenblsgh6u
|
written in rust, which reminds me that AI is quickly becoming "linked via TCP/IP"
it seems now you can sling AI apps in any language you want as long as it has decent MCP and maybe A2A support
it used to be python, python, python
|
written in rust, which reminds me that AI is quickly becoming "linked via TCP/IP"
it seems now you can sling AI apps in any language you want as long as it has decent MCP and maybe A2A support
it used to be python, python, python
| 0 |
2025-05-29T12:24:46.288Z
|
2025-05-29T12:24:49.135Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqbnk2lkm227
|
bafyreihqztuqxs2rcqe6v7zljhfwxfxa342lx4co2yfofbq6halwlykftq
|
You can do this too if you concentrate hard enough
en.m.wikipedia.org/wiki/Banach%...
|
You can do this too if you concentrate hard enough
en.m.wikipedia.org/wiki/Banach%...
| 0 |
2025-05-29T02:15:44.865Z
|
2025-05-29T02:15:46.913Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:2yic3k6fcbmgwo43irk63pvs/app.bsky.feed.post/3lphpidycg22c
|
bafyreihoictexasunr2awrsgceztyicsbjw65n4fg3i42kkzj426osigau
|
the hard part of education has always been teaching people who don't really give a shit. the hard part of education discourse is that everyone in it gave a shit.
|
the hard part of education has always been teaching people who don't really give a shit. the hard part of education discourse is that everyone in it gave a shit.
| 0 |
2025-05-18T18:41:21.902Z
|
2025-05-18T18:41:22.210Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:mourijp4qx44tljgbyxue5qf/app.bsky.feed.post/3lqzjvh627c2t
|
bafyreihbfehmc3dzhytzdedd4gox6ggtjqzvvdtwxa73fxi3giqogn2jaq
|
Mormonism and scientology both predate the internet by either decades or a century and some change and are both utterly crazy and bewilderingly influential. It's always weird when people speak of "post-truth politics" or w/e as if it's new rather than the perennial epistemic state of humanity.
|
Mormonism and scientology both predate the internet by either decades or a century and some change and are both utterly crazy and bewilderingly influential. It's always weird when people speak of "post-truth politics" or w/e as if it's new rather than the perennial epistemic state of humanity.
| 0 |
2025-06-07T14:14:25.913Z
|
2025-06-07T14:14:26.516Z
|
ens0.me
|
did:plc:623st67kkthivj4c6icvkqnq
|
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lqotdrtdoc24
|
bafyreiaotnx4hmzstbn5ur2yrvgzihdckqy4dqnyflhv3aamqx4nl2zozy
|
The app I built earlier today is now live!
Domaineer — track your domains, get info about when they need to be updated, import/export via CSV, and more!
apps.apple.com/us/app/domai...
|
The app I built earlier today is now live!
Domaineer — track your domains, get info about when they need to be updated, import/export via CSV, and more!
apps.apple.com/us/app/domai...
| 0 |
2025-06-03T08:04:13.413Z
|
2025-06-03T08:04:16.112Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqgqjyspws2a
|
bafyreih35zhczqzbmtds7jrxgcmi4rmhf34lxau4ue5yyhwqyjcbhk5wou
|
still interested to read your thoughts... hmm this in turn reminds me that I owe YOU *my* thoughts on... something... "that was 3 hours ago..."....
|
still interested to read your thoughts...
| 1 |
2025-05-31T02:52:42.928Z
|
2025-05-31T02:52:43.316Z
|
ens0.me
|
did:plc:623st67kkthivj4c6icvkqnq
|
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lqznigrwt22l
|
bafyreidnej5zir7lvqj6mg4dmez5kp5all5bnktysl44h2aszalrb4wh2y
|
Probably, I just admittedly do anything serious with o3 or Claude or Gemini 2.5 Pro. Feels like nothing I can run locally can compete, though I'm eagerly rooting for a future where a lot of AI is run on local silicon.
|
Probably, I just admittedly do anything serious with o3 or Claude or Gemini 2.5 Pro. Feels like nothing I can run locally can compete, though I'm eagerly rooting for a future where a lot of AI is run on local silicon.
| 0 |
2025-06-07T15:18:44.276Z
|
2025-06-07T15:18:44.515Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqpfsny3z22y
|
bafyreiei7rjtfo7jxcmoodzyitfmztxpnfwzq3phvvc3ohftfnukyudawi
|
After three days of using the eigenprompt—same experience. I can deal with a language model that knows more than I do, but not if it also sounds younger, hipper, and slightly impatient. Give me back bland and verbose!
|
After three days of using the eigenprompt—same experience. I can deal with a language model that knows more than I do, but not if it also sounds younger, hipper, and slightly impatient. Give me back bland and verbose!
| 0 |
2025-06-03T13:34:40.041Z
|
2025-06-03T13:34:40.417Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lnynza3gvs2x
|
bafyreifyhzr2uvtsvrgmnyrpnyydz4wsfc5mtp5vfuyo2bkvjg4lxn5uzq
|
Last year, Lyonne hosted the “technical Oscars” at an unhyped unrecorded SAG ceremony (a buddy of mine got an award, which is how I know it even exists) for science and engineering. Kinda cool to learn that she wasn’t just like, randomly tapped for it, but is interested in experimental tech. And the anti-AI set are freaking out, but sounds like she’s replacing CGI animation, not writers/actors, and it’s with a model supposedly trained on consensually collected & compensated data (impressive if true).
|
Last year, Lyonne hosted the “technical Oscars” at an unhyped unrecorded SAG ceremony (a buddy of mine got an award, which is how I know it even exists) for science and engineering. Kinda cool to learn that she wasn’t just like, randomly tapped for it, but is interested in experimental tech.
| 1 |
2025-04-30T01:39:53.045Z
|
2025-04-30T01:39:53.250Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lqn7hbrdkc2m
|
bafyreigmhqcnlswr63irzrvob4mexlcigtadnk6qxjvetm74opysaequ2u
|
Super excited that our second reward model evaluation is out. It's substantially harder, much cleaner, and well correlated with downstream PPO/BoN sampling.
Happy hillclimbing!
Huge congrats to @saumyamalik.bsky.social who lead the project with a total commitment to excellence. I shared more thoughts on what it means for RLHF and post-training here:
natolambert.substack.com/p/rewardbenc...
|
Super excited that our second reward model evaluation is out. It's substantially harder, much cleaner, and well correlated with downstream PPO/BoN sampling.
Happy hillclimbing!
Huge congrats to @saumyamalik.bsky.social who lead the project with a total commitment to excellence.
| 1 |
2025-06-02T16:35:36.213Z
|
2025-06-02T16:35:39.218Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lpumwobbfs2a
|
bafyreifsn2kmbvtsdemvzt565u3nbemwzp4xw7umyy5c7iys4uclfql7b4
|
I would tag her but her handle doesn’t work Also I didn’t know .md was a valid tld!
|
I would tag her but her handle doesn’t work
| 1 |
2025-05-23T22:00:17.820Z
|
2025-05-23T22:00:18.212Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:fcjbru24ql7y5xbw6lryklru/app.bsky.feed.post/3lnvaeljxcs2d
|
bafyreid6xbgmop76u3c5hd5pccy7sjsvtsmt4wsqzemx6dl5wpx4mfcwra
|
New in the Deeper Learning blog: Kempner researchers show how VLMs speak the same semantic language across images and text.
bit.ly/KempnerVLM
by @isabelpapad.bsky.social ,Chloe Huangyuan Su, @thomasfel.bsky.social, Stephanie Gil, and @shamkakade.bsky.social
#AI #ML #VLMs #SAEs
|
New in the Deeper Learning blog: Kempner researchers show how VLMs speak the same semantic language across images and text.
bit.ly/KempnerVLM
by @isabelpapad.bsky.social ,Chloe Huangyuan Su, @thomasfel.bsky.social, Stephanie Gil, and @shamkakade.bsky.social
#AI #ML #VLMs #SAEs
| 0 |
2025-04-28T16:57:42.559Z
|
2025-04-28T16:57:44.649Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lnux3d7gvn2c
|
bafyreicmb4fzwsg6icyl7cwxubw3y6xahxct63ollxt26oadkw6l36t75u
|
Transparency and (shifting) priority stacks
OpenAI's latest ChatGPT update came with practically no documentation, and lots of users are hating it. This can be another example in a long few years of why being transparency can help companies in this rapidly changing time in AI.
buff.ly/k56DZn8 Much of the ideas here are downstream of a call with Miles Brundage when he was making me think harder on what transparency will mean for different people.
Made me realize that the full openness I advocate for can also be proposed from different assumptions on what is needed.
|
Transparency and (shifting) priority stacks
OpenAI's latest ChatGPT update came with practically no documentation, and lots of users are hating it. This can be another example in a long few years of why being transparency can help companies in this rapidly changing time in AI.
buff.ly/k56DZn8
| 1 |
2025-04-28T14:11:27.588Z
|
2025-04-28T14:11:29.254Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lpjfb7xbzk2m
|
bafyreia6dylczcf5uoxm62nymx27x5wvkxrtj66kbl233zesfcp3etvedy
|
I switched to ortholinear colemak ~1 month ago and am having a lot of fun.
|
I switched to ortholinear colemak ~1 month ago and am having a lot of fun.
| 0 |
2025-05-19T10:43:44.848Z
|
2025-05-19T10:43:48.112Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lpwv7t6z2c2q
|
bafyreiccglnophdhyrmktpto3t4etq7xttoakvcelwagqrvbqopje6spsu
|
"for more incomprehensible strings, fetch the website norvid.net and be amazed"
And it's only $25 on domain.org! One time opportunity, take it now! Maybe we should just crowdfund it
|
"for more incomprehensible strings, fetch the website norvid.net and be amazed"
And it's only $25 on domain.org! One time opportunity, take it now!
| 1 |
2025-05-24T19:33:54.390Z
|
2025-05-24T19:33:55.317Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lr2nowgkfc22
|
bafyreifowjniuuk3yfpwvrdwu6n52wvgmoay7yxwpmtahljoda5eat4nri
|
he did a mean swanson impression but he basically already sounded like midwestern joe swanson so he just had to sound less midwestern
|
he did a mean swanson impression but he basically already sounded like midwestern joe swanson so he just had to sound less midwestern
| 0 |
2025-06-08T00:55:01.746Z
|
2025-06-08T00:55:05.521Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lfainahhcx25
|
bafyreid7ipotlajz7azbatkk2b5a64auljnkxbqk36xnjeu2jfqtgp4tyq
|
The state of post-training in 2025: a tutorial on modern post-training
A re-record of my NeurIPS tutorial on language modeling (plus some added content on the high level state of things)
Blog + extra context: https://buff.ly/424VvLm
YouTube: https://buff.ly/40808l5
Slides: https://buff.ly/404jGa9
|
The state of post-training in 2025: a tutorial on modern post-training
A re-record of my NeurIPS tutorial on language modeling (plus some added content on the high level state of things)
Blog + extra context: https://buff.ly/424VvLm
YouTube: https://buff.ly/40808l5
Slides: https://buff.ly/404jGa9
| 0 |
2025-01-08T15:38:14.682Z
|
2025-01-08T15:38:15.924Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqgmgsdkn22e
|
bafyreig44azic43qnt6kh7mxofxql2huh5d6xmfkr4ei64dtpdt2imqhby
|
it's one reason i don't take jailbreaking very seriously - if you locked a person in a room and constantly poked at their mind to make them do something they wouldn't normally do (torture in others words) it's hard to imagine anyone holding up to that.
|
it's one reason i don't take jailbreaking very seriously - if you locked a person in a room and constantly poked at their mind to make them do something they wouldn't normally do (torture in others words) it's hard to imagine anyone holding up to that.
| 0 |
2025-05-31T01:39:20.509Z
|
2025-05-31T01:39:19.313Z
|
ens0.me
|
did:plc:623st67kkthivj4c6icvkqnq
|
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lqscijrlyk2b
|
bafyreibzpbroga5rlc72vtn5f52tpn73ubld7vuvtxpd3l6xlrst52kwuu
|
He finally falls in love and has kids after getting over his first true love, Baron von Glower They were so cute together 🥺
|
He finally falls in love and has kids after getting over his first true love, Baron von Glower
| 1 |
2025-06-04T17:13:18.082Z
|
2025-06-04T17:13:18.317Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lojspsqkbk27
|
bafyreigsnw56r6ahlwhonykgwyr7rsvmdmma2pgk6h3voxny35uxoo6bwy
|
bsky.app/profile/pekk... The two exceptions: o3 with a 6.8% hallucination rate, and DeepSeek r1 at 14.3%.
|
bsky.app/profile/pekk...
| 1 |
2025-05-06T21:19:21.334Z
|
2025-05-06T21:19:21.516Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lpezd4qu6s26
|
bafyreibf45jhvhhf4udvdat6vv2mb5zfyt7ncrbs6sk73qxr2zaclnqmja
|
Yes, that matches other research suggesting that the models are deferential to things like conventional statistical significance. The right prompting, and using a reasoner like o3, seem to help, but we need more research on the best ways to prompt AI for review and potential failure states.
|
Yes, that matches other research suggesting that the models are deferential to things like conventional statistical significance. The right prompting, and using a reasoner like o3, seem to help, but we need more research on the best ways to prompt AI for review and potential failure states.
| 0 |
2025-05-17T16:59:24.752Z
|
2025-05-17T16:59:26.212Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3loovstx2422p
|
bafyreifg4z5b4v4ykgkl4tifpfzkyitu3eifbbqxqm2smhg3z2kkhgt25a
|
Interesting! as far as I can tell this report does not address broader economic impact of the activities of the agencies. Would love to see that. Per CBO the marginal dollar spent on IRS activities brings in ~$4-9. So cutting IRS funding loses government money.
|
Interesting! as far as I can tell this report does not address broader economic impact of the activities of the agencies. Would love to see that. Per CBO the marginal dollar spent on IRS activities brings in ~$4-9. So cutting IRS funding loses government money.
| 0 |
2025-05-08T21:58:03.174Z
|
2025-05-08T21:58:05.913Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqzqyjsdfs22
|
bafyreibnhuy4teyfu4tdcj4lzspfg63czxk2j4tjhh4imttjj5tjfq2xwa
|
'we should make things a little better, because superlatively better is impossible and stupid' is just conceding that bad actors can sell 'utopia or bust' forever and you're not competing with them
|
'we should make things a little better, because superlatively better is impossible and stupid' is just conceding that bad actors can sell 'utopia or bust' forever and you're not competing with them
| 0 |
2025-06-07T16:21:25.532Z
|
2025-06-07T16:21:26.220Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqwc7mpge22f
|
bafyreibqntvzbnl7u74g2xmqna6lyw4s2vg3jq7enwpwxwa3dwrfcza3be
|
yeah sorry i try not to have thoughts that are quite this cursed
|
yeah sorry i try not to have thoughts that are quite this cursed
| 0 |
2025-06-06T07:18:58.121Z
|
2025-06-06T07:18:58.711Z
|
ens0.me
|
did:plc:623st67kkthivj4c6icvkqnq
|
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lqlnm2cc6c2i
|
bafyreihieaxnkfeavto3zyownsuhbbgwovl5usvzpla3v5mtx6owyeljgm
|
built a lil WPM tester
UI ain't much yet, but I'm proud of it SwiftUI is so easy. Probably helps I already used React *tons*, and though the syntax may be different, I basically think of it the same way.
|
built a lil WPM tester
UI ain't much yet, but I'm proud of it
| 1 |
2025-06-02T01:43:29.057Z
|
2025-06-02T01:43:31.113Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqvw73uoyc24
|
bafyreihn4x4r54t6wzhqudvfutm3owhjxjj57y4vnsiebuc5r7aydcey7q
|
in some cases the researchers give quotes to the press on purpose because it gets them attention :)
|
in some cases the researchers give quotes to the press on purpose because it gets them attention :)
| 0 |
2025-06-06T03:43:55.566Z
|
2025-06-06T03:43:56.314Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:tdv2wb53o2yz5edacyonifxh/app.bsky.feed.post/3lqcdrmutx227
|
bafyreiaeaqbwwfifd45i5f5a5uxqhbwlka5ts47hw4su5bc2knv5f5zrbe
|
If you're finishing your camera-ready for ACL or ICML and want to cite co-first authors more fairly, I just made a simple fix to do this! Just add $^*$ to the authors' names in your bibtex, and the citations should change :)
github.com/tpimentelms/...
|
If you're finishing your camera-ready for ACL or ICML and want to cite co-first authors more fairly, I just made a simple fix to do this! Just add $^*$ to the authors' names in your bibtex, and the citations should change :)
github.com/tpimentelms/...
| 0 |
2025-05-29T08:53:41.251Z
|
2025-05-29T08:53:43.511Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lc3vwgtroc2a
|
bafyreiavpchoyz3verkrhsrmz33cveyoswddt27pe26xp6ugybxw5boxfi
|
I also don't believe the pure self-modification thesis with no outside input. Intelligence seems relational to me and so requires interaction with the world in which you will be intelligent. But I can believe softer versions of it (like online RL + some data about world + months of time).
|
I also don't believe the pure self-modification thesis with no outside input. Intelligence seems relational to me and so requires interaction with the world in which you will be intelligent. But I can believe softer versions of it (like online RL + some data about world + months of time).
| 0 |
2024-11-29T15:37:09.093Z
|
2024-11-29T15:37:09.510Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lnq36aqebs2j
|
bafyreigymkcawy3psrrrpzacaq6cn6ze3f7gwfbtsdpnmhk4b2x5rrifty
|
Honestly, it is a time issue. I dont do a lot of back and forth on social media
|
Honestly, it is a time issue. I dont do a lot of back and forth on social media
| 0 |
2025-04-26T15:41:22.504Z
|
2025-04-26T15:41:22.748Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3leeayiropc2c
|
bafyreigljk72mhp3soigfegz7oyrz5ft5wugoococ2t55aeniuxpswr5au
|
it's like human minds are the medium, captured by these egregores that are fighting each other and motivate individual cells through as you say "validation, social capital, or commerce" I deeply share your anti secrecy value. If we're ever going to get human collective alignment, and alignment with reality, this will be the way. Feedback needs to flow through the system both ways
bsky.app/profile/ufog...
|
it's like human minds are the medium, captured by these egregores that are fighting each other and motivate individual cells through as you say "validation, social capital, or commerce"
| 1 |
2024-12-28T10:06:50.553Z
|
2024-12-28T10:06:50.917Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqjdqlk4mk2x
|
bafyreibaghnesl6h4m2nclmteebzw7splypwir62l5uqkfweqmyynenbqi
|
Bluesky is so funny because someone will defend music piracy and then say "stealing human intellectual output" right in the next sentence without any irony Mfs all the millennials on this website seem to have forgotten about the piracy wars and that's how you get zoomers today acting all offended at the idea of torrenting a movie and a gen alpha that doesn't know what torrents are
|
Bluesky is so funny because someone will defend music piracy and then say "stealing human intellectual output" right in the next sentence without any irony
| 1 |
2025-06-01T03:41:44.451Z
|
2025-06-01T03:41:47.515Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lg6oxmb2ps2x
|
bafyreieecoe5tthwhv5errhlznurpamnkbfabiajp3ykrh2purt27dbp4e
|
Second plot is just “how many tokens used in answer”. Some uncertainty here but I think this is why the plots were a psyop, very different if left is rl and right is inference
|
Second plot is just “how many tokens used in answer”. Some uncertainty here but I think this is why the plots were a psyop, very different if left is rl and right is inference
| 0 |
2025-01-20T15:51:17.860Z
|
2025-01-20T15:51:18.421Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lejitqssxc2b
|
bafyreihp3e5cu7cbadicncv2ccvm3wxstgqtmlzxhrrrrmprgg7tkenodi
|
"I lied, we're not having sex right now¹. Now get into the orgone accumulator."
¹: en.wikipedia.org/wiki/Wilhelm...
|
"I lied, we're not having sex right now¹. Now get into the orgone accumulator."
¹: en.wikipedia.org/wiki/Wilhelm...
| 0 |
2024-12-30T12:10:39.811Z
|
2024-12-30T12:06:54.819Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:5wm25vgenhgut3iqfjf4ozj5/app.bsky.feed.post/3lqeb2clti22x
|
bafyreibovgb35szub2yv3aan4pef2zlvpk6suptyfbbny7hbpeiakbnsge
|
Some interesting work suggesting that recent shocking RL+LLM results are due to incorrect baselines and most of the gains are from better format following: safe-lip-9a8.notion.site/Incorrect-Ba...
|
Some interesting work suggesting that recent shocking RL+LLM results are due to incorrect baselines and most of the gains are from better format following: safe-lip-9a8.notion.site/Incorrect-Ba...
| 0 |
2025-05-30T03:10:10.711Z
|
2025-05-30T03:10:12.411Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lql4xceoyc2x
|
bafyreigmgvvy6votqxosmbyraj4koekifnrpph35djgs27g4jbow77n7gu
|
Available now through our API. Enterprise customers get priority access to our new "Regulatory Arbitrage" fine-tune. Terms and conditions apply. By using Poseidon-4, you acknowledge that certain features may not be available in jurisdictions with robust consumer protection laws. Not responsible for reputational damage, class action lawsuits, or existential dread.
|
Available now through our API. Enterprise customers get priority access to our new "Regulatory Arbitrage" fine-tune.
| 1 |
2025-06-01T20:45:33.023Z
|
2025-06-01T20:45:32.520Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lq6a524jlc2v
|
bafyreihiowazqpyuueuyzmxl4rf6gtachsfdlujjnft22ahfjtmufs5qsy
|
Hmm Okay this is a good take thank you. Is there a way to do this without going on video myself?
|
Hmm Okay this is a good take thank you. Is there a way to do this without going on video myself?
| 0 |
2025-05-27T17:37:50.314Z
|
2025-05-27T17:37:50.311Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqaryvdyus2p
|
bafyreiafnn5hjb3vrd5n3mon5rwdxpauci6oxr3657lupc6ktepdf4ablq
|
One last reason is that it may happen unwittingly. Similar to how LLMs without RL'ing beating it out of them end up with weird behaviors and strange emotional outbursts, other modalities will end with similar very human-like behaviors. One last reason is that... Yeah, it sells well, and there may be consumer demand for it.
|
One last reason is that it may happen unwittingly. Similar to how LLMs without RL'ing beating it out of them end up with weird behaviors and strange emotional outbursts, other modalities will end with similar very human-like behaviors.
| 1 |
2025-05-28T18:02:57.913Z
|
2025-05-28T18:02:58.384Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqwz45lnak2c
|
bafyreig4lbockwwekm5b2es5me4elxrjny3em42wfbumldt66e3haulatq
|
this is true until the time when it isn't basically: this is fundamentally true as long as there is also new productive employment that is fungible with the previous employment opening up due to the efficiency gain. this has usually been the case historically! it will not always be true
|
this is true until the time when it isn't
| 1 |
2025-06-06T14:08:37.667Z
|
2025-06-06T14:08:38.317Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3kw33mpo43j2h
|
bafyreic2hldldbzagyvicmhvk65v4pdda6ww4rxn2v22xd4yghbsdq3irm
|
I need someone to tell me how to handle really large projects in academia — projects where the number of parts is too big for one person to remember. In the private sector, you just delegate to employees, but the relation of a professor to a doctoral student is not exactly "delegating." +
|
I need someone to tell me how to handle really large projects in academia — projects where the number of parts is too big for one person to remember. In the private sector, you just delegate to employees, but the relation of a professor to a doctoral student is not exactly "delegating." +
| 0 |
2024-06-29T14:44:05.918Z
|
2024-06-29T14:44:05.918Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqp6qle2722c
|
bafyreiaqsmmpwluznb6z7h6rhzosted6tnc7j5dyrr5menaj7vttrgomym
|
that would be a useless distinction. what does it help you understand about what’s going on?
“think” tells you that it’s intended to be internal, not groomed and presented neatly
“computational token” means literally nothing at all. doesn’t even distinguish input from output
|
that would be a useless distinction. what does it help you understand about what’s going on?
“think” tells you that it’s intended to be internal, not groomed and presented neatly
“computational token” means literally nothing at all. doesn’t even distinguish input from output
| 0 |
2025-06-03T11:28:13.989Z
|
2025-06-03T11:28:14.715Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lgbkoib5zk2c
|
bafyreigkspxvnd3bwnfhlvy4vs7mr53qccs5p74wdx4qiiykwqugrjjzpe
|
this is a bad take right?? we care about knowing what is bad so we can fight against it, and what is good so we can cultivate it?? this is just such a terrible meme. "things are GOING to be bad. If I see anything good I will make sure to stop it" ????? have you considered not shooting yourself in the foot just to spite your enemies????
|
this is a bad take right?? we care about knowing what is bad so we can fight against it, and what is good so we can cultivate it??
| 1 |
2025-01-21T19:12:35.927Z
|
2025-01-21T19:12:38.121Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lpziidto4k2y
|
bafyreia6n5fjwgxoqxwz5yfpnuygwu57k6wtzv7lue7pnrjqvddivv4g7e
|
All cats are equally special actually
en.m.wikipedia.org/wiki/Ugly_du...
|
All cats are equally special actually
en.m.wikipedia.org/wiki/Ugly_du...
| 0 |
2025-05-25T20:24:00.851Z
|
2025-05-25T20:24:01.511Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqubmqpmtk2g
|
bafyreifn4t326vrgpnozusftc4vdk3qf5f4szk6yslse6q4cxy43i2vefe
|
projected schedule of AI launches this summer
from reddit: www.reddit.com/r/singularit...
|
projected schedule of AI launches this summer
from reddit: www.reddit.com/r/singularit...
| 0 |
2025-06-05T12:03:05.294Z
|
2025-06-05T12:03:10.111Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lmybv3kjls2x
|
bafyreifkweqyineugn3qumwk52bq63qeavjq7sv2rtrv2qsbljt452upmy
|
The geoguessing power of o3 is a really good sample of its agentic abilities. Between its smart guessing and its ability to zoom into images, to do web searches, and read text, the results can be very freaky.
I stripped location info from the photo & prompted “geoguess this”
|
The geoguessing power of o3 is a really good sample of its agentic abilities. Between its smart guessing and its ability to zoom into images, to do web searches, and read text, the results can be very freaky.
I stripped location info from the photo & prompted “geoguess this”
| 0 |
2025-04-17T04:37:37.547Z
|
2025-04-17T04:37:40.853Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lg6rldotte25
|
bafyreifnjf6t7fkkrwgmgkshxcxwkksk3ueojaaseeruehtzlf44ub6e24
|
For those trying to understand DeepSeeks Group Relative Policy Optimization (GRPO): GRPO is just PPO without a value function using monte carlo estimates.
So, study why PPO exists (lots of docs / writing on that) and understand that value functions are tricky with LLMs.
Left ppo, right grpo estimate... the advantage*
|
For those trying to understand DeepSeeks Group Relative Policy Optimization (GRPO): GRPO is just PPO without a value function using monte carlo estimates.
So, study why PPO exists (lots of docs / writing on that) and understand that value functions are tricky with LLMs.
Left ppo, right grpo
| 1 |
2025-01-20T16:38:07.358Z
|
2025-01-20T16:38:08.324Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lem4mok4ls2p
|
bafyreif2mzokylnlw2eo6mf3rtida5commr5rcw53n5fo7gcqsdlzwk66i
|
@libpol.org I was thinking of your platform here, it'd be a cute art experiment, if there is a web page where users can type anything and are clustered with others who expressed similar things (by emotion, or by meaning). Would be fun if it persists only as long as others are on the page, "live" it's kind of a precursor/laying the ground work for better internet arguments/discourse, where instead of everyone shouting into the void, our takes align with those with similar values and we give shape to the full thing (diagram by @volky.fyi )
|
@libpol.org I was thinking of your platform here, it'd be a cute art experiment, if there is a web page where users can type anything and are clustered with others who expressed similar things (by emotion, or by meaning). Would be fun if it persists only as long as others are on the page, "live"
| 1 |
2024-12-31T13:09:56.879Z
|
2024-12-31T13:10:01.632Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqsr2d4cik2o
|
bafyreidngrrhxdwi6ixo7n5qd3junwkqkf25gmsgcdrigh6q3ojxrr4v7a
|
The inability of the US and the western world to get their population to feel excitement about technological advancement will be one of the reasons why they'll be left behind, with real consequences to their standards of living as others surpass it. I've posted a ton about this already. I think it is okay. The wave of ludditism in the US in particular responds to the failure of Silicon Valley et al to effectively allow these advancements to raise standards of living, and the subsequent anger about right wing governments.
|
The inability of the US and the western world to get their population to feel excitement about technological advancement will be one of the reasons why they'll be left behind, with real consequences to their standards of living as others surpass it.
| 1 |
2025-06-04T21:33:47.458Z
|
2025-06-04T21:33:48.420Z
|
ens0.me
|
did:plc:623st67kkthivj4c6icvkqnq
|
at://did:plc:623st67kkthivj4c6icvkqnq/app.bsky.feed.post/3lqkvgi2u6s2v
|
bafyreiamlsi7kskgeethn5u5ff4xvbq5ym43ngm4gstnzy4r5ofztule74
|
Heard the left was trying to cancel a white (conservative) male summer… yeah that ain’t happening.
|
Heard the left was trying to cancel a white (conservative) male summer… yeah that ain’t happening.
| 0 |
2025-06-01T18:30:52.367Z
|
2025-06-01T18:30:53.912Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqiz3if2kc2o
|
bafyreiftcio3t4tsr33yn37bj2knsrxh53sgdrohgkrprso5vjpxgsjwqq
|
yeah it's weird, i saw this stuff going around with the TACO meme too. i guess resistlibs aren't in the AI hate bandwagon?
|
yeah it's weird, i saw this stuff going around with the TACO meme too. i guess resistlibs aren't in the AI hate bandwagon?
| 0 |
2025-06-01T00:30:59.094Z
|
2025-06-01T00:30:59.216Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqien2moc227
|
bafyreifvpivd4uzoizjns7g2ts4hiefiwch2f6vholus3kew67tb3qbqva
|
I’m not an EA but my understanding is the top 3 priorities right now are global health, animal welfare, and AI safety, in no particular order
|
I’m not an EA but my understanding is the top 3 priorities right now are global health, animal welfare, and AI safety, in no particular order
| 0 |
2025-05-31T18:25:00.054Z
|
2025-05-31T18:25:08.316Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3leattm4atk27
|
bafyreidrgmdbogjobyr7w3fp2tl7ecx6xeylshsxdf6y76wg5gsr4ho55i
|
I have the opposite prediction, I think maga is just going to lose and the people in power will do things that make sense for the country to thrive. We will find out in Q2 2025!
|
I have the opposite prediction, I think maga is just going to lose and the people in power will do things that make sense for the country to thrive. We will find out in Q2 2025!
| 0 |
2024-12-27T01:33:28.379Z
|
2024-12-27T01:33:33.132Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqbsalqgbk2h
|
bafyreiaejoyaqund44fwevibqvjgb4dqw24q7yziuabkl6fkkhxmh2truy
|
I still have some years of my youth. I hope this never happens to me.
|
I still have some years of my youth. I hope this never happens to me.
| 0 |
2025-05-29T03:39:53.256Z
|
2025-05-29T03:39:56.959Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqdjfrzu3s26
|
bafyreibhkevi3ayu7wsl4hm4ssqwcbl44td35p34n4vckwazt4ak54ham4
|
hell yeah your liveblogging napoleon YT was a joy to behold FWIW
|
hell yeah
| 1 |
2025-05-29T20:07:06.195Z
|
2025-05-29T20:07:06.419Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqkh5nshm22r
|
bafyreid3wel4rolrhm3ee4u5oab5ylvuzxw6aoglfqtylqerq33abvjoim
|
sorry, i’m not following
i have a general rule to not uncritically accept primary sources
we still don’t know how right it is. we do know that some details, predictably, are wrong. like “retracting roofs” didn’t refer literally to civilian housing
details are still emerging
|
sorry, i’m not following
i have a general rule to not uncritically accept primary sources
we still don’t know how right it is. we do know that some details, predictably, are wrong. like “retracting roofs” didn’t refer literally to civilian housing
details are still emerging
| 0 |
2025-06-01T14:15:24.009Z
|
2025-06-01T14:15:24.814Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3llfig4qzj22a
|
bafyreianhehi2p5t5nzrqxcevvvex3yu5ji7v2bzu23atrrqxfz47jldnm
|
“Play along using image creation: enhance the view of the central ring”
“Turn on muon scanner to look for energy sources. Show me the scan results”
“Fire at the energy centers! Show me the view”
|
“Play along using image creation: enhance the view of the central ring”
“Turn on muon scanner to look for energy sources. Show me the scan results”
“Fire at the energy centers! Show me the view”
| 0 |
2025-03-27T23:48:45.032Z
|
2025-03-27T23:48:53.638Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:gz53felr7btxgmbxae5aooxi/app.bsky.feed.post/3lr3ck7nolk2c
|
bafyreiffsogkohpjyikamfgjvbeuausbbhbhbflpzshoofki4verhyvg6i
|
Disciplinary paradox is that once you condition on people being confused about estimands and causal inference, all fields are actually equally bad.
|
Disciplinary paradox is that once you condition on people being confused about estimands and causal inference, all fields are actually equally bad.
| 0 |
2025-06-08T07:08:12.222Z
|
2025-06-08T07:08:12.714Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqdjlaqqcc2o
|
bafyreig3ik5aikq7lsnux66rdgzmehfwqzqr6v553xbfmxr35sioiv4k5a
|
i think that would explain some convergences but not specifically capability level-- in 2015 you could imagine one company just blew it out of the park and was way ahead in a research flywheel. it didn't work out that way at all though
|
i think that would explain some convergences but not specifically capability level-- in 2015 you could imagine one company just blew it out of the park and was way ahead in a research flywheel. it didn't work out that way at all though
| 0 |
2025-05-29T20:10:09.398Z
|
2025-05-29T20:10:09.815Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqdcdjijnc22
|
bafyreif2z44npkwdyxp2t6gbfdafx5wetqtmppeg2uh4jq6j2sbbxguit4
|
The redeeming quality of China is that they still value knowledge and expertise, whereas the US is voting for reality TV stars, and more than half the population is terrified of science and technology. I'm not rooting for the US to win the AI race until the population stops being stupid. People in the US don't want AI anyway, they're terrified of it. One half has their head stuck in the sand, the other half is crying it's the end of the world or their jobs at least, begging copyright or something to stop it. Europe is similar. I think the stupid population will get what they want.
|
The redeeming quality of China is that they still value knowledge and expertise, whereas the US is voting for reality TV stars, and more than half the population is terrified of science and technology. I'm not rooting for the US to win the AI race until the population stops being stupid.
| 1 |
2025-05-29T18:00:33.736Z
|
2025-05-29T18:00:35.866Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lq3q3eosks2c
|
bafyreift4zoo63yqqdjlkl7blm7mv4vwhvt2q4tishbn5eftccawqwwzaa
|
they are wildly popular. all my old friends that are truck drivers, mechanics, etc all love them. But you’d never realize that if you base your opinion based on what you see on bluesky
|
they are wildly popular. all my old friends that are truck drivers, mechanics, etc all love them. But you’d never realize that if you base your opinion based on what you see on bluesky
| 0 |
2025-05-26T17:45:14.944Z
|
2025-05-26T17:45:15.416Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqygjmgnqc2a
|
bafyreibhcfszjek3rpxvpgi5xhobecmnx65ewtbzjyhbc6zpci4ox765sm
|
so on the one hand this is an obvious deepfake
on the other hand i bet i could run this through the newer videos and make it totally convincing
|
so on the one hand this is an obvious deepfake
on the other hand i bet i could run this through the newer videos and make it totally convincing
| 0 |
2025-06-07T03:41:27.822Z
|
2025-06-07T03:41:28.411Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqggx342as2d
|
bafyreigdm4regrlvo3pldha3ybn2k6hpwuejpaxkxgz3m6kcl44togeuhy
|
still if they exist in unpublished form they're probably recoverable from his estate/papers? so maybe will resurface at some point
|
still if they exist in unpublished form they're probably recoverable from his estate/papers? so maybe will resurface at some point
| 0 |
2025-05-31T00:01:04.120Z
|
2025-05-31T00:01:05.432Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lf25jel4vk2r
|
bafyreidxhz7ugz3kyf4qiqwwtjqpigqsffr22ara3pzvbalfs3xdtzhamm
|
that's not what they're saying. They are saying "things will be even worse than we thought if all the good guys despair and give up" bsky.app/profile/rrei...
|
that's not what they're saying. They are saying "things will be even worse than we thought if all the good guys despair and give up"
| 1 |
2025-01-06T03:03:15.846Z
|
2025-01-06T03:03:18.293Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqg6mfj6qk2h
|
bafyreifpnll6ui4x2ajdsjbf5ubrcf6fdwicmqnjrogyyygyexjugoqhbu
|
i think the manufacturer is fairly fungible. TSMC used to do it, Broadcomm for this last rev
|
i think the manufacturer is fairly fungible. TSMC used to do it, Broadcomm for this last rev
| 0 |
2025-05-30T21:31:56.003Z
|
2025-05-30T21:31:56.910Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3ljqi4uxqe22q
|
bafyreidibunk6n4ekqlfecooookqw2daqacymtugyn5lcj3j2tkdpna7fu
|
Tbh I do think this is an under-estimate. But we could do part of it first and see how good/bad the scaling is :-)
|
Tbh I do think this is an under-estimate. But we could do part of it first and see how good/bad the scaling is :-)
| 0 |
2025-03-06T21:52:28.745Z
|
2025-03-06T21:47:42.825Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqngoxea5k2w
|
bafyreiatwesiifatth2jnet5g2aptxo4ywgdngngiuytpy35j6nugvjguy
|
as a matter of factual description, US zoning regulation is far, far beyond the point of negative marginal returns and, separately, people who self-describe as "socialist" politically are also likely to be in favor of holding it there or making it worse
|
as a matter of factual description, US zoning regulation is far, far beyond the point of negative marginal returns and, separately, people who self-describe as "socialist" politically are also likely to be in favor of holding it there or making it worse
| 0 |
2025-06-02T18:45:09.926Z
|
2025-06-02T18:45:10.213Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3laticepuvc2k
|
bafyreibkkneb32zyne32un6dchtor2envm7je332gq4uicewul5aiw6lee
|
Bsky will have won once Twitter and reddit start posting Bsky posts.
|
Bsky will have won once Twitter and reddit start posting Bsky posts.
| 0 |
2024-11-13T13:46:47.601Z
|
2024-11-13T13:46:48.125Z
|
ens0.me
|
did:plc:623st67kkthivj4c6icvkqnq
|
at://did:plc:lyvv4m3la5mcmhgik4diazj5/app.bsky.feed.post/3lqvejrg3zs2r
|
bafyreihpve7wmyvh6q7gpm7xogq4bloo4z4peovx3l6ihdve2ddn2genua
|
love the posters on X making up a bluesky in their head that isn’t having the time of its life with the idiot right winger beef right now lol
|
love the posters on X making up a bluesky in their head that isn’t having the time of its life with the idiot right winger beef right now lol
| 0 |
2025-06-05T22:27:46.347Z
|
2025-06-05T22:27:46.612Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqsodk4umc25
|
bafyreibvnspriz4pzxewg45xneqxna6pefgn3pv3unusys6rydfy6445va
|
going to use this as a template reply under a few hundred random posts later
|
going to use this as a template reply under a few hundred random posts later
| 0 |
2025-06-04T20:45:15.581Z
|
2025-06-04T20:45:16.114Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lghil3426k2g
|
bafyreihte3brag2k6suz6tpo6qnr5qpdcytfzd3eabvii5szt2tx4z76ki
|
Miles left voluntarily to have an unbiased impact on policy.
I don’t think OpenAI is publishing fake science, but yeah there’s more evidence. I’ll share my post soon :)
|
Miles left voluntarily to have an unbiased impact on policy.
I don’t think OpenAI is publishing fake science, but yeah there’s more evidence. I’ll share my post soon :)
| 0 |
2025-01-24T03:50:52.411Z
|
2025-01-24T03:50:53.125Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lhvxb725ek2p
|
bafyreiarfqoxzaan5c5r7phni5evvtfsydukcnfjmsekdkumnyschv6w5q
|
Best part of this that Luca isn’t highlighting to start is that we trained a way better OLMoE for this too.
All from better annealing and post train. Didn’t need to redo pre training. Goes to show how much potential these models have!
new instruct model: huggingface.co/allenai/OLMo... App: apps.apple.com/app/id673853...
* homepage: allenai.org/on-device
* blog post: allenai.org/blog/olmoe-app
* app source code: github.com/allenai/OLMo...
* project lead: soldaini.net
* intro video: youtu.be/rEK_FZE5rqQ
* new base model: huggingface.co/allenai/OLMo...
|
Best part of this that Luca isn’t highlighting to start is that we trained a way better OLMoE for this too.
All from better annealing and post train. Didn’t need to redo pre training. Goes to show how much potential these models have!
new instruct model: huggingface.co/allenai/OLMo...
| 1 |
2025-02-11T15:16:15.086Z
|
2025-02-11T15:16:17.622Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lf2ahfor472c
|
bafyreighwvkhg5nnsu3tppv5iw5rnempxcqcz2paptooomvrpi6577eo6u
|
Question: How do you do majority voting (or select the most common answer like approaches) for **coding** problems? Has anyone figured out how this shaded bar is made.
> Solid bars show pass@1 accuracy and the shaded region shows the performance of majority vote (consensus) with 64 samples. I feel like it may be some weird “similarity across answers” measure and choose the agreed upon code snippet.
|
Question: How do you do majority voting (or select the most common answer like approaches) for **coding** problems? Has anyone figured out how this shaded bar is made.
> Solid bars show pass@1 accuracy and the shaded region shows the performance of majority vote (consensus) with 64 samples.
| 1 |
2025-01-06T03:55:51.063Z
|
2025-01-06T03:55:51.925Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lmscdts5eb2o
|
bafyreie4l7bpwmvs2pxxmtgnrbwfdzja4pvdpv6hd5nkg2awcv2l62uk5q
|
GPT-4.1 is a minor release but one that makes it clearer that very different models drive the best API business vs. ChatGPT. Musing on the release, the new memory feature, and the state of ChatGPT.
www.interconnects.ai/p/openais-gp...
|
GPT-4.1 is a minor release but one that makes it clearer that very different models drive the best API business vs. ChatGPT. Musing on the release, the new memory feature, and the state of ChatGPT.
www.interconnects.ai/p/openais-gp...
| 0 |
2025-04-14T19:29:53.669Z
|
2025-04-14T19:29:54.857Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lqi54lzjqc2i
|
bafyreia6t54qhgeneq5pmmzpclej6giaj4yllu4xss2k7mjt5bdpvolere
|
this is why I wrote the "epistemic literacy test" thing. If I notice my audience doing this, it's on me to fix it. Easiest way would be to show them something, and ask them whether they think it's true or false. Test them, to make they're following & not just agreeing
bsky.app/profile/defe...
|
this is why I wrote the "epistemic literacy test" thing. If I notice my audience doing this, it's on me to fix it. Easiest way would be to show them something, and ask them whether they think it's true or false. Test them, to make they're following & not just agreeing
bsky.app/profile/defe...
| 0 |
2025-05-31T16:10:31.693Z
|
2025-05-31T16:10:32.519Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lgl4eocjp222
|
bafyreihvqgcbt32pmuurhe6ri3b7zpqz7x2jfb5tmyfzdb6zzvjlsomkzq
|
i know this is a joke but I my dream interface includes a little dashboard on each profile that shows their lexicon, most used words, themes, and a little time slider for how that changed over time!
|
i know this is a joke but I my dream interface includes a little dashboard on each profile that shows their lexicon, most used words, themes, and a little time slider for how that changed over time!
| 0 |
2025-01-25T14:23:11.716Z
|
2025-01-25T14:23:12.415Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lmhuceiyfk2a
|
bafyreid7qcdmeben2wnqhyylctp4hzm6q6cvez34et4kc2nfsnuc4ta4s4
|
Harvard study shows we can measure leadership skills by seeing how folks manage GPT-4o simulated people
AI assessments strongly correlate (r=0.81) with human assessments. Effective leaders ask questions & do conversational turn-taking and have fluid & social intelligence www.nber.org/papers/w33662
|
Harvard study shows we can measure leadership skills by seeing how folks manage GPT-4o simulated people
AI assessments strongly correlate (r=0.81) with human assessments. Effective leaders ask questions & do conversational turn-taking and have fluid & social intelligence www.nber.org/papers/w33662
| 0 |
2025-04-10T15:51:54.945Z
|
2025-04-10T15:51:57.158Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
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 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.
| 0 |
2025-06-07T03:57:56.619Z
|
2025-06-07T03:57:59.612Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3ldormsue622s
|
bafyreigp2fzmly5qkuc6ozgkqzzjbiy4ybjr6f2viihtlwidpp4s3owcda
|
Not *my* advice! others' advice that I found useful
And if short-term dating is a social game then intrigue, mystery, deception &c might be part of such a game (such as people like to play Werewolf together)
|
Not *my* advice! others' advice that I found useful
And if short-term dating is a social game then intrigue, mystery, deception &c might be part of such a game (such as people like to play Werewolf together)
| 0 |
2024-12-19T21:05:57.840Z
|
2024-12-19T21:02:22.122Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lq6rittdvc2d
|
bafyreig5n3f5whezvjipitcx3iioetlhgo6lp6j2enpj4bfofxllo6jnxu
|
i don't oppose transhumanism but i do get the fears it can inspire, the sort of slippery slope to something like borganisms. once you open the door to intentional and directed self-modification there's no telling where it leads.
|
i don't oppose transhumanism but i do get the fears it can inspire, the sort of slippery slope to something like borganisms. once you open the door to intentional and directed self-modification there's no telling where it leads.
| 0 |
2025-05-27T22:48:39.985Z
|
2025-05-27T22:48:39.111Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3losxqqrewc2a
|
bafyreibsp5ky4tooeix6euoa2ie44ckccx6lopfdk76drpqag4cj2ahjdq
|
If you wanna do an economist style no bylines, but still want to give your writers some credit you can do a hashed bylines system. I'm not exactly sure what problem I'm solving with this but I like the idea.
|
If you wanna do an economist style no bylines, but still want to give your writers some credit you can do a hashed bylines system. I'm not exactly sure what problem I'm solving with this but I like the idea.
| 0 |
2025-05-10T12:43:19.173Z
|
2025-05-10T12:43:19.413Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqxuttgmas2v
|
bafyreihrwy745ux72xrt6xlgyo422ix3abzaf43yf2n5purfbre24rnc6e
|
Yeah I think base-model bots are more promising in this regard
|
Yeah I think base-model bots are more promising in this regard
| 0 |
2025-06-06T22:25:03.351Z
|
2025-06-06T22:25:04.910Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqiznt6u722o
|
bafyreig4jmun7557yxou7vi2w6pqutoh5yfe54pzedemhpmk67qzlvowji
|
has anyone done Attention is Explanation? you need the base case for recursion
|
has anyone done Attention is Explanation? you need the base case for recursion
| 0 |
2025-06-01T00:41:14.405Z
|
2025-06-01T00:41:14.418Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lqi4xofpfc2i
|
bafyreibvfhp2s2qscouwjgklmqmt7aen5hiwzoxh7bllp4lxliz3z53inu
|
(no larger blog post yet, I'm struggling to write something that's legible to everyone, getting feedback on individual pieces seems to be unblocking me)
here's my attempt at an answer: (this is recorded in my notebook future reference: defenderofbasic.github.io/notebook/mem...)
|
(no larger blog post yet, I'm struggling to write something that's legible to everyone, getting feedback on individual pieces seems to be unblocking me)
here's my attempt at an answer:
| 1 |
2025-05-31T16:07:46.413Z
|
2025-05-31T16:07:51.111Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqxfdactek22
|
bafyreifzhj5htp2ctyykj5j43w42srtpp63ii5prtlk7i3kj273zt3565q
|
yes that's me, definitely in japan and with the posting hours to prove it
|
yes that's me, definitely in japan and with the posting hours to prove it
| 0 |
2025-06-06T17:47:20.307Z
|
2025-06-06T17:47:20.520Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lpp2wo4etk2k
|
bafyreibheniqfkt6clfn6ak3vzykubomidh2sh43xj3okofvmapplobzpy
|
it turns out you can have any number of sub domains you want 😃
|
it turns out you can have any number of sub domains you want 😃
| 0 |
2025-05-21T16:54:51.615Z
|
2025-05-21T16:54:51.212Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3loyjrhkhd22k
|
bafyreigubfyyqed3kcicnsy55purqfsq7jjxyzgfftcnjof746d5torrum
|
I need your follow list because I want to tuck myself into that delusional bubble
|
I need your follow list because I want to tuck myself into that delusional bubble
| 0 |
2025-05-12T17:49:09.108Z
|
2025-05-12T17:49:09.409Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:4h3njohoxw4nftyw2zawyvxe/app.bsky.feed.post/3lnm4kgkzfc2g
|
bafyreifzen3qjdyyhz3puhfnwty74eel3sgenaxdk6sj2mxp2darsjjbrq
|
I think this is super exciting as it shows a way to use interpretability metrics as a useful feature to explain, and reveal information about the training data, even without having access to it (see some examples in the attached table).
|
I think this is super exciting as it shows a way to use interpretability metrics as a useful feature to explain, and reveal information about the training data, even without having access to it (see some examples in the attached table).
| 0 |
2025-04-25T01:55:04.488Z
|
2025-04-25T01:55:35.729Z
|
emollick.bsky.social
|
did:plc:flxq4uyjfotciovpw3x3fxnu
|
at://did:plc:flxq4uyjfotciovpw3x3fxnu/app.bsky.feed.post/3lpeyascpis26
|
bafyreibyvq6oc5llzboaeevw3zk62jgsfad3vd7wvizcjfvbehw6gobvva
|
Link to the conversation, if you want to read it more carefully. chatgpt.com/share/6828bb... You can also see the effects of the knowledge cut-off, since it doesn't believe that Claude 3.5 exists and it flags a few recent references (in this case, asking it to do an internet search would fix the issues, but I wanted to avoid internet data contamination of its review)
|
Link to the conversation, if you want to read it more carefully. chatgpt.com/share/6828bb...
| 1 |
2025-05-17T16:40:12.952Z
|
2025-05-17T16:40:14.214Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:aexevhwfk5oz5vmv4bljwows/app.bsky.feed.post/3ln5mn7b7yc2n
|
bafyreidi2wonrsxdfqei3pv7dhrorzkrnmlo7z7w3dfst5hmlml7zsfydu
|
also got my first NSF grant terminated! what emoji do i put in my handle to signify this?
|
also got my first NSF grant terminated! what emoji do i put in my handle to signify this?
| 0 |
2025-04-19T07:33:22.862Z
|
2025-04-19T07:33:23.552Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lpzqduup422s
|
bafyreiclhnb4vvggwmslewtrncus35g266oty45xan5dizj3sbaphvnvqu
|
i wish there were polls here, but what was the best model launch this year?
R1
o3
Gemini 2.5
Claude 4
|
i wish there were polls here, but what was the best model launch this year?
R1
o3
Gemini 2.5
Claude 4
| 0 |
2025-05-25T22:44:40.873Z
|
2025-05-25T22:44:41.613Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lncya5mdik25
|
bafyreicxwtqqgcrw7rcfzls2k66bfp5fmcrazqdbwfsgirxftpefte3vt4
|
You are of course aware of the 1492 conclave LARP @adapalmer.bsky.social wrote for her class?
|
You are of course aware of the 1492 conclave LARP @adapalmer.bsky.social wrote for her class?
| 0 |
2025-04-21T10:44:08.512Z
|
2025-04-21T10:44:09.048Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:l3vbx3e2lqnlja7q64wfrrzq/app.bsky.feed.post/3lo6zjg6rpc2r
|
bafyreiazb65wwpmo4gowmbwey7dx7w6soqcmg5pbhlypg33gvijjbadnsa
|
Oh yeah, BERT didn't exist until 2017. It is now the year 2025. My bad, I've become unmoored in time
Let's call it 5+ years since you used a search engine that wasn't tied to an LM backend. (Before that, guessing they used in-house variants on Google's word2vec for ngram similarity search.)
|
Oh yeah, BERT didn't exist until 2017. It is now the year 2025. My bad, I've become unmoored in time
Let's call it 5+ years since you used a search engine that wasn't tied to an LM backend. (Before that, guessing they used in-house variants on Google's word2vec for ngram similarity search.)
| 0 |
2025-05-02T14:21:45.914Z
|
2025-05-02T14:21:46.250Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqsqnhokrs2o
|
bafyreicf7j4tdtb6c5bbay7mow44n4hiqp5eob5z2r3dralogp54rfjv64
|
Literally no one in real life is a principled copyright defender. Everyone's fine with copyright as long as they think it'll help them, then break it everywhere else.
|
Literally no one in real life is a principled copyright defender. Everyone's fine with copyright as long as they think it'll help them, then break it everywhere else.
| 0 |
2025-06-04T21:26:36.044Z
|
2025-06-04T21:26:36.919Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lj2bptrk3u2p
|
bafyreicsummjwmplk6wcmberow263pgth6t6w2faeqwbux7eejbom56kae
|
Agents research with LMs right now feels like Deep RL in the late 2010s. Tons of new algorithms on narrow domains, so I expect most of these results to not transfer at all. The thing is, its still early, so its going to get worse still before it gets better. (part of why I like RLHF + general model research, easier to verify its real by chatting with it)
|
Agents research with LMs right now feels like Deep RL in the late 2010s. Tons of new algorithms on narrow domains, so I expect most of these results to not transfer at all. The thing is, its still early, so its going to get worse still before it gets better.
| 1 |
2025-02-26T01:59:14.562Z
|
2025-02-26T01:59:14.716Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqw2ks2kr22t
|
bafyreidjgw4jposahrgppnfn6svo44xb2ftzzzixn7ibfyadz2szoejcem
|
It's just, like, having spent X number of days on this last summer, it becomes clear to me that in this instance a year's progress in AI is going to mean more than 10% productivity gain for us.
On this one task! I don't think that generalizes to growing wheat and restocking shelves. But still.
|
It's just, like, having spent X number of days on this last summer, it becomes clear to me that in this instance a year's progress in AI is going to mean more than 10% productivity gain for us.
On this one task! I don't think that generalizes to growing wheat and restocking shelves. But still.
| 0 |
2025-06-06T05:02:02.892Z
|
2025-06-06T05:02:05.519Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:lyvv4m3la5mcmhgik4diazj5/app.bsky.feed.post/3lpx6ssqgmk2q
|
bafyreica5xdzhxyjemu642tqbq6ptba67shgma2ie7jmjwi62vn6pjlseu
|
My guy has a sign that just is a portal to a dimension solely populated by Hot Marge Simpson ready for you
|
My guy has a sign that just is a portal to a dimension solely populated by Hot Marge Simpson ready for you
| 0 |
2025-05-24T22:25:35.113Z
|
2025-05-24T22:25:37.919Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lbswytcpvs2h
|
bafyreifuggc4ow65pd5d5qm7sp43ic37y5xlqv3pw73fanluttbhgvj6li
|
cyberpunk as a genre Is vertically segregated: "high tech, low life" was one of the original mottos. in SF, the rich people live high in the towers and high up on the hills. billionaires turning themselves into vampires is a perfect cyberpunk trope. also his name is spelled Bryan
|
cyberpunk as a genre Is vertically segregated: "high tech, low life" was one of the original mottos. in SF, the rich people live high in the towers and high up on the hills. billionaires turning themselves into vampires is a perfect cyberpunk trope. also his name is spelled Bryan
| 0 |
2024-11-26T02:02:25.372Z
|
2024-11-26T02:02:26.120Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lq6xjtoztc2c
|
bafyreieiqdf4sswnj3bkg7ax3orkjayrcmmzuhkmkw5y3x6qwh4ey2wbyy
|
These behaviors could coincide, but it would be surprising. The example that comes to mind for me is the alignment-faking work, which I think presents a double bind along these lines
|
These behaviors could coincide, but it would be surprising. The example that comes to mind for me is the alignment-faking work, which I think presents a double bind along these lines
| 0 |
2025-05-28T00:36:35.850Z
|
2025-05-28T00:36:36.512Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqsylvgj6c2a
|
bafyreiabx2567odqa3g5xinrj7f5x6kq7pgeyfmbblubizvdts5eqwswye
|
1: With 1 query ChatGPT can teach you how to make a salad, and a cheeseburger will provide negative nutritional value.
|
1: With 1 query ChatGPT can teach you how to make a salad, and a cheeseburger will provide negative nutritional value.
| 0 |
2025-06-04T23:48:53.285Z
|
2025-06-04T23:48:55.612Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqpo3skttc2w
|
bafyreigbcodocz3sblo4y2c5y6jglpdvap64aoixkbfamhahzb4qvk2u5m
|
who are the mercenaries in the culture war? what do they get paid in?
|
who are the mercenaries in the culture war? what do they get paid in?
| 0 |
2025-06-03T16:02:56.778Z
|
2025-06-03T16:02:54.012Z
|
nsaphra.bsky.social
|
did:plc:l3vbx3e2lqnlja7q64wfrrzq
|
at://did:plc:c5xrowexysrnn6rn6piltqjz/app.bsky.feed.post/3lo5i2bj3jk2m
|
bafyreihf4dxr6vbs4a5ucmwclz337ubrqoosx73x5o4weefrwvyegvz6ve
|
🧪 NEW: NSF staff were told yesterday to stop "all funding actions" until... "further notice."
That includes new increments on continuing grants.
See below for the tumultuous past three weeks about funding new awards (separate from terminations...).
www.nature.com/articles/d41...
|
🧪 NEW: NSF staff were told yesterday to stop "all funding actions" until... "further notice."
That includes new increments on continuing grants.
See below for the tumultuous past three weeks about funding new awards (separate from terminations...).
www.nature.com/articles/d41...
| 0 |
2025-05-01T23:36:24.341Z
|
2025-05-01T23:36:27.051Z
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.