handle
stringclasses 12
values | did
stringclasses 12
values | uri
stringlengths 70
70
| cid
stringlengths 59
59
| text
stringlengths 41
614
| originalText
stringlengths 3
301
| replyCount
int64 0
3
| createdAt
stringdate 2023-06-01 10:40:18
2025-06-07 21:48:17
| indexedAt
stringdate 2023-06-01 10:40:18
2025-06-07 21:48:18
|
|---|---|---|---|---|---|---|---|---|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqqd5dqa6227
|
bafyreiglwbnofusyuy6bmuj3jvg7g4yr433n72fcqla3otm5zmt6lyctva
|
Eigenprompt is really grating to me personally but I might try a less stylized variant of it to see if I can get the same effect I also think format constraints like "lowercase only" might nerf models a little, but it's probably negligible at this point.
|
Eigenprompt is really grating to me personally but I might try a less stylized variant of it to see if I can get the same effect
| 1 |
2025-06-03T22:19:36.913Z
|
2025-06-03T22:19:37.911Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqzj7lpxnc27
|
bafyreicse3nnhnjnbayfgfohjsxjhsjtukatvbo6434lyp5nrmivb2q6t4
|
The discourse is already bad, but the Internet has hosted all kinds of strange misinformation for 3 decades. I expect that for anything important, it'll be fact-checked pretty quickly. People who want to believe things will continue to do so. Also no one will believe that Trump sucked Elon's toes. As funny as that'd be.
|
The discourse is already bad, but the Internet has hosted all kinds of strange misinformation for 3 decades. I expect that for anything important, it'll be fact-checked pretty quickly. People who want to believe things will continue to do so.
| 1 |
2025-06-07T14:02:12.498Z
|
2025-06-07T14:02:12.817Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqx4nkv4ns2t
|
bafyreihtpdd3dwv5nwautouqjvdzivzin4nj26fsoa3kalhabvbaigcnxy
|
i used them to write an entire readme. and i rewrote the entire readme! but it saved me the first draft and the 'going and reading fifteen code files to make sure i am not missing anything' steps
|
i used them to write an entire readme. and i rewrote the entire readme! but it saved me the first draft and the 'going and reading fifteen code files to make sure i am not missing anything' steps
| 0 |
2025-06-06T15:12:03.260Z
|
2025-06-06T15:12:05.215Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lqqaxvs7jc27
|
bafyreib23zasizabzgs4sgj6al2osuvzlwetklq7h2cf5g6dntz74tnnmy
|
Yeah seems true though I'm putting more & more stuff into my system prompt to get Claude to be less sycophantic
I should play around more with Gemini, that was actually pretty fun in its bluntness
|
Yeah seems true though I'm putting more & more stuff into my system prompt to get Claude to be less sycophantic
I should play around more with Gemini, that was actually pretty fun in its bluntness
| 0 |
2025-06-03T21:40:47.040Z
|
2025-06-03T21:40:48.813Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3ljqk5me6vc2w
|
bafyreibdsgrq7ktnbenlqwv3v33elnpgr55dmouayqmy5ir4mzdjl47bcm
|
I could do this, bumping it up a bit on the priority list
|
I could do this, bumping it up a bit on the priority list
| 0 |
2025-03-06T22:28:40.753Z
|
2025-03-06T22:23:55.118Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqyobrzrzk23
|
bafyreiggst7m45snfoyjqbgmxrjlh2cxoje7kza4oguif6db5qlhrk6jvu
|
if i dredge my memory maybe i can find where the third one is but probably not
|
if i dredge my memory maybe i can find where the third one is but probably not
| 0 |
2025-06-07T06:00:15.191Z
|
2025-06-07T06:00:15.912Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqx6f2ltsk2n
|
bafyreidgfdhcjch7kitnajpsa5mhibppqhfzwlbpcvaoqa6b4py6w67noq
|
lmao fair. i accept my fate if it escapes containment and am willing to defend that it is true at least half of the time
|
lmao fair. i accept my fate if it escapes containment and am willing to defend that it is true at least half of the time
| 0 |
2025-06-06T15:43:05.227Z
|
2025-06-06T15:43:05.917Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lnqzxkjxvs2r
|
bafyreiavwrnitv27ukqgsy2bpgywsyw2ujxxxyj64rhgjkdo7rjfktzphm
|
More @ niplav.site/nootropics.h...
I have been informed that I should've been taking oil with Vitamin D₃ during this experiment, so the results here are likely not as strong as I'd hoped. Also it's a fairly small dose. Still, I'm happy with it. I also know this kinda contradicts received wisdom on Vitamin D₃. I'm may try to figure out what's going on there, but more likely I'll just shrug my shoulders and go "okay, whatever, metabolism is weird". Send meta-analyses
|
More @ niplav.site/nootropics.h...
I have been informed that I should've been taking oil with Vitamin D₃ during this experiment, so the results here are likely not as strong as I'd hoped. Also it's a fairly small dose. Still, I'm happy with it.
| 1 |
2025-04-27T00:52:23.876Z
|
2025-04-27T00:52:25.049Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqn3mmpomk2r
|
bafyreib5xcvkr5tx3apokqlpntxv4jrc7rw36tx3plda4pf3pdfyabnt4y
|
i feel like this isn’t obvious until you’ve tried it
there’s a huge difference between one search and an agent making repeated searches, each iteration asking more refined questions
|
i feel like this isn’t obvious until you’ve tried it
there’s a huge difference between one search and an agent making repeated searches, each iteration asking more refined questions
| 0 |
2025-06-02T15:27:00.499Z
|
2025-06-02T15:27:01.220Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqklitrqnk2r
|
bafyreibakt3qt4zbtxsxnbjognzhugbzlr5r6fcn64e4pjpc7eaqahs4im
|
yes! it needs all the data before it makes up its mind (exploration). you can go extremely deep with it, but follow-up questions are really just for clarification. it’s not gonna change its mind
|
yes! it needs all the data before it makes up its mind (exploration). you can go extremely deep with it, but follow-up questions are really just for clarification. it’s not gonna change its mind
| 0 |
2025-06-01T15:33:14.343Z
|
2025-06-01T15:33:15.014Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqutuea3fk2b
|
bafyreicq537g52rdzio5ch2wnybxruwwdp6owh7lxmu5hkgrbpfekacfci
|
>looking for a new llm
>ask the ai bros if their llm is goodharted or eldritch
>they dont understand
>pull out illustrated diagram explaing what is goodharted and what is eldritch
>they laugh and says “it’s a good llm sir”
>buy a subscription
>its goodharted
|
>looking for a new llm
>ask the ai bros if their llm is goodharted or eldritch
>they dont understand
>pull out illustrated diagram explaing what is goodharted and what is eldritch
>they laugh and says “it’s a good llm sir”
>buy a subscription
>its goodharted
| 0 |
2025-06-05T17:29:28.007Z
|
2025-06-05T17:29:28.512Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lnds3xo7sk2z
|
bafyreigewtvzeweiwdnamvmvgn6bu4tezhc4quh2llfmpvjfudu3xvgnha
|
That's my best current guess I hope you prove me wrong
|
That's my best current guess
| 1 |
2025-04-21T18:27:05.355Z
|
2025-04-21T18:27:08.158Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqxgmhxwgc2w
|
bafyreie2wvg7fuw3uwhw3jbtvxeze2hnwmmatkyiqcphoan55uneg4lq7a
|
I think that’s true, but the fact that it *can* look so different is still surprising I think
|
I think that’s true, but the fact that it *can* look so different is still surprising I think
| 0 |
2025-06-06T18:10:24.069Z
|
2025-06-06T18:10:24.410Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lkqdtuqg4i2g
|
bafyreias7vg2vkzkmvtl5bxhheqxfu2eik7bkthom2vozqempa6lczlqwe
|
Something I've been doing background research (journalism? nah) on for a while! What makes leading AI libs go brr
Managing frontier model training organizations (or teams)
How do the frontier labs consistently train great models? How can they fail?
|
Something I've been doing background research (journalism? nah) on for a while! What makes leading AI libs go brr
Managing frontier model training organizations (or teams)
How do the frontier labs consistently train great models? How can they fail?
| 0 |
2025-03-19T14:01:02.655Z
|
2025-03-19T14:01:05.528Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqxhc65r732p
|
bafyreiapfvaq2k5pw742ioaave55lpu3fhk2b5jhuc3qpylvvfxpjxiyti
|
Auren and Seren do a convincing job of it, but I think they require so much scaffolding and they seem kind of limited in how the illusion works. Like they respond 2 or 3 messages at a time, and aren't very variable in their behavior.
|
Auren and Seren do a convincing job of it, but I think they require so much scaffolding and they seem kind of limited in how the illusion works. Like they respond 2 or 3 messages at a time, and aren't very variable in their behavior.
| 0 |
2025-06-06T18:22:31.973Z
|
2025-06-06T18:22:33.312Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3ll55nqhicc2e
|
bafyreiadzdhmvbggctcfi3fqwww5n5zaphiotvg2qjswkae7et2ceaigay
|
yep, just a context doc, predates Claude by 3 years. I won't reveal the claude convo because obvious privacy worries are obvious :-D
|
yep, just a context doc, predates Claude by 3 years. I won't reveal the claude convo because obvious privacy worries are obvious :-D
| 0 |
2025-03-24T16:14:51.508Z
|
2025-03-24T16:09:38.319Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lquvzmrnik2p
|
bafyreihtslhrwvlon4wwd5h2nj5egot3voptboj4ofiwaoc474l5ggoowu
|
That's why it's great! You have to have a bunch of dumb people to succeed as a social network site. I'm doing my part certainly
|
That's why it's great! You have to have a bunch of dumb people to succeed as a social network site. I'm doing my part certainly
| 0 |
2025-06-05T18:08:12.227Z
|
2025-06-05T18:08:13.415Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:i4kytxgsu3yfsrt2ml3o7tgq/app.bsky.feed.post/3lqn77j443n2v
|
bafyreie5szx4ictg23ye6hpe4sfzi5l4zxrm2qwwpxfqb53ijffvjp3otu
|
RewardBench 2 is here! We took a long time to learn from our first reward model evaluation tool to make one that is substantially harder and more correlated with both downstream RLHF and inference-time scaling.
|
RewardBench 2 is here! We took a long time to learn from our first reward model evaluation tool to make one that is substantially harder and more correlated with both downstream RLHF and inference-time scaling.
| 0 |
2025-06-02T16:31:15.449Z
|
2025-06-02T16:31:16.314Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:ecc3o7r4wpmoiejq3rcx3m22/app.bsky.feed.post/3lqyfhccif22g
|
bafyreigmogsnafm6vsmnjmqsjxdlevixczah4ehvhq3ogxegrkk5fpsh5u
|
The first iteration of DeepPast's modular AI architecture for digital humanities research (developed with Donghyeok Choi, Eric Chow, Michael Chung, Chi Man Kwong, Yumeng Hou, and Miguel Escobar Varela).
To be unveiled at the World Congress on Sinology 2025 closing keynote in Shenzhen this afternoon.
|
The first iteration of DeepPast's modular AI architecture for digital humanities research (developed with Donghyeok Choi, Eric Chow, Michael Chung, Chi Man Kwong, Yumeng Hou, and Miguel Escobar Varela).
To be unveiled at the World Congress on Sinology 2025 closing keynote in Shenzhen this afternoon.
| 0 |
2025-06-07T03:22:16.348Z
|
2025-06-07T03:22:21.016Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:rvlyeda73kxm7l2weegk73pa/app.bsky.feed.post/3lqnzd677is2p
|
bafyreic7nchj4hphaobneyvbp5q3jmtuqhssg3d2wudoemu3phbni2v7v4
|
Wow LLMs are pretty good at reading and explaining TLA+ error traces (and syntax errors). This could be real big for learners
Claude also did a great job of converting a spec from using 1 server to using N servers, an annoying bit of finicky boilerplate it handled well
|
Wow LLMs are pretty good at reading and explaining TLA+ error traces (and syntax errors). This could be real big for learners
Claude also did a great job of converting a spec from using 1 server to using N servers, an annoying bit of finicky boilerplate it handled well
| 0 |
2025-06-03T00:18:35.544Z
|
2025-06-03T00:18:36.810Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lnuyrj6cvi2u
|
bafyreif5ksvmpc6upzetlrfx7op6zat6re2g7qnjbwadokgdjsijt6tijm
|
With Qwen launching a 0.6B model with their v3 lineup, we should keep in mind what the smallest model that is robust in chat and capable of some inference-time scaling with RL is. This size going down over time will determine functionality for a lot of on-device and interesting applications. 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.
|
With Qwen launching a 0.6B model with their v3 lineup, we should keep in mind what the smallest model that is robust in chat and capable of some inference-time scaling with RL is. This size going down over time will determine functionality for a lot of on-device and interesting applications.
| 1 |
2025-04-28T14:41:46.332Z
|
2025-04-28T14:41:46.453Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lodze3jj4k2f
|
bafyreic6l3n6kfwjmlmsv2d4jatun2qrngdfhtdepdllxpdbdaabd55rem
|
Have you wondered why millions of people and thousands of researchers deem these models useful and you don't? I think it is very important for our future that people understand that these models are not just another parlor trick, but will become a significant part of our economy in the next 5 years
|
Have you wondered why millions of people and thousands of researchers deem these models useful and you don't? I think it is very important for our future that people understand that these models are not just another parlor trick, but will become a significant part of our economy in the next 5 years
| 0 |
2025-05-04T14:02:05.651Z
|
2025-05-04T14:02:06.351Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqxcqeu6ls2k
|
bafyreibpocoyv72aeegsmcc3n5gtn6tnd4hirvjwihhfhlsdlpnqtiq3tq
|
on the upload image button any video i select is grayed out. maybe it has a max time of 5 seconds or something?
|
on the upload image button any video i select is grayed out. maybe it has a max time of 5 seconds or something?
| 0 |
2025-06-06T17:01:00.052Z
|
2025-06-06T17:01:00.421Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqkcmlfhf22r
|
bafyreicgxwrakav7u2jseuayn2ppomxaygpzibvnjcsxdotnuwdrtmemxe
|
i think that’s a weak point of LLMs, they can look at photos, even videos, but not enough to really validate this stuff
|
i think that’s a weak point of LLMs, they can look at photos, even videos, but not enough to really validate this stuff
| 0 |
2025-06-01T12:54:16.092Z
|
2025-06-01T12:54:17.112Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqv6rr4rok26
|
bafyreifqrwfgf6om6h2mumqrngmoylmtzirpu7azjwokoepjwvkyqer6ou
|
who is that specifically though. one part is 'pro israel/anti campus protests', that doesnt really cover the NSF though
|
who is that specifically though. one part is 'pro israel/anti campus protests', that doesnt really cover the NSF though
| 0 |
2025-06-05T20:44:52.027Z
|
2025-06-05T20:44:52.318Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqee4cydk22k
|
bafyreidsxujyabj2isby4xlrvult3tll5hvucjlvijlzbumfbjbto3nnzy
|
I read this as embodied dev and was extremely curious
|
I read this as embodied dev and was extremely curious
| 0 |
2025-05-30T04:04:59.457Z
|
2025-05-30T04:05:00.112Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqx44sydw22t
|
bafyreigoxke35rvt6jxwbcow6lbthqrl2257q7px6ffaqbidyojwtrzfem
|
i mean i think the schtick is bad and doing it once after name searching yourself is pretty discrediting
|
i mean i think the schtick is bad and doing it once after name searching yourself is pretty discrediting
| 0 |
2025-06-06T15:02:41.328Z
|
2025-06-06T15:02:42.111Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:2p7a3z4uzfpoepv3aa36ykjq/app.bsky.feed.post/3lquovdguhc2y
|
bafyreig4k6hdrnex6nwfbmw5i55q3w6yysbfeii246ztahkyxqud3dhnoi
|
there's been a few times I tried pull-starting my social media browsing in bluesky and it was unsuccessful
this latest time in the past month or so? absolutely no downside, feels great.
I think some kind of critical mass of poasters arose here that the engine can stay alive with
|
there's been a few times I tried pull-starting my social media browsing in bluesky and it was unsuccessful
this latest time in the past month or so? absolutely no downside, feels great.
I think some kind of critical mass of poasters arose here that the engine can stay alive with
| 0 |
2025-06-05T16:00:32.025Z
|
2025-06-05T16:00:32.712Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lq65gjoxqi2i
|
bafyreierkezxgusz6jkfe4ju3ftjzs3nu5jmvbkpjwhl2hniwb6owx2tma
|
What this looks like on different models, and with different rewards, is that a lot of the recent toy RLVR papers like training on one example are exploiting something weird about qwen.
Read more about the paper from the leads: buff.ly/ZWx83Mj
But there's more to say about RLVR... We need to get into the space where academics are SCALING RL and not just playing with weird behaviors (investigative work) in the models. We are nearly there and can make the plunge soon!
That's the argument I make in this post summarizing the situation: buff.ly/IT5lWEc
|
What this looks like on different models, and with different rewards, is that a lot of the recent toy RLVR papers like training on one example are exploiting something weird about qwen.
Read more about the paper from the leads: buff.ly/ZWx83Mj
But there's more to say about RLVR...
| 1 |
2025-05-27T16:49:27.397Z
|
2025-05-27T16:49:27.920Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqrrhlgnos2s
|
bafyreiar4n5dawfn5awizrr5wrkw7vjf27fflken3ctfhmfgdzofmbcrzu
|
I’m not saying that statistical ML is as safe as mop tops, or even that mop tops themselves were safe! Some perceived risks are real.
I’m just interested in how poorly we perceive our own era, esp on social media. The intensity of daily controversy is more salient here than decade-scale change.
|
I’m not saying that statistical ML is as safe as mop tops, or even that mop tops themselves were safe! Some perceived risks are real.
I’m just interested in how poorly we perceive our own era, esp on social media. The intensity of daily controversy is more salient here than decade-scale change.
| 0 |
2025-06-04T12:08:30.811Z
|
2025-06-04T12:08:33.453Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lmrxzqncn52f
|
bafyreia2l4lrxff4olcwryjyfbf7tzeucwru3honwpnxdvpzxzpwn5am5i
|
If we say overfitting to lmsys is bad (ie high score = cooked model), then we can't also say a low score is bad (claude historically has been low).
But, in context of everything else w llama 4, is 👀
|
If we say overfitting to lmsys is bad (ie high score = cooked model), then we can't also say a low score is bad (claude historically has been low).
But, in context of everything else w llama 4, is 👀
| 0 |
2025-04-14T16:25:17.962Z
|
2025-04-14T16:25:18.553Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqsys5szek2a
|
bafyreiabk26j7prgtpwjqwnwko33wdyuvonyb3oia3otjohkodhkgxrvfy
|
Generic recipe on Google will make me scroll through 10 pages of a personal story just to make me see ads and that'll make me lose calories, providing negative nutritional value.
|
Generic recipe on Google will make me scroll through 10 pages of a personal story just to make me see ads and that'll make me lose calories, providing negative nutritional value.
| 0 |
2025-06-04T23:52:23.411Z
|
2025-06-04T23:52:24.112Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqg37cmkxk2t
|
bafyreifsicycvrgxyvnj4qilfz3w4yb367mnd7pf6dhajrnulregy6ti3u
|
It’s going to be okay because we’re going to make it okay
|
It’s going to be okay because we’re going to make it okay
| 0 |
2025-05-30T20:30:55.534Z
|
2025-05-30T20:30:56.210Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lqsemvx2jc2k
|
bafyreidjim77skxsjj5d5nxnqdkoohbrbjnevdh4xzktfv7yrmdkno5sau
|
What is a network state, if not a dinner party perservering?
|
What is a network state, if not a dinner party perservering?
| 0 |
2025-06-04T17:51:32.546Z
|
2025-06-04T17:51:32.914Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lqi5u3bovc2s
|
bafyreihh7amdsu4inpfimeuao6ljwkwjlpod5rohrkfph4ov57qk2eq2ri
|
When you say "OpenAI is featuring" you mean "this user-generated content reached the front page" not "San Altman deliberately created this" right San
|
When you say "OpenAI is featuring" you mean "this user-generated content reached the front page" not "San Altman deliberately created this" right
| 1 |
2025-05-31T16:23:39.441Z
|
2025-05-31T16:23:39.813Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lezbwh2s6c2d
|
bafyreie7tk4nl4qjssosipttumevntq5j5avwumcwmake2bxhkl2advsc4
|
"hope is frail yet hard to kill" is the most true & beautiful thing I've read all week 💜
|
"hope is frail yet hard to kill" is the most true & beautiful thing I've read all week 💜
| 0 |
2025-01-05T18:49:29.893Z
|
2025-01-05T18:49:28.528Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqvvdchxt224
|
bafyreidyzlwrnvq77nwme6ulbwemnu3zzmhhmhtl4l5mmqb52bzqjeflxm
|
we got exactly three ubi checks and then stopped for no good reason
|
we got exactly three ubi checks and then stopped for no good reason
| 0 |
2025-06-06T03:28:22.965Z
|
2025-06-06T03:28:23.518Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqypi24wlk2h
|
bafyreihfpwzy5mu7zjvd4divjubic4uzo7lhoa7dfmkk5u6b2krhgiusoi
|
shhhhh we don't want them to figure it out
|
shhhhh we don't want them to figure it out
| 0 |
2025-06-07T06:21:38.751Z
|
2025-06-07T06:21:39.513Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqqgpktkdk2d
|
bafyreiggspoysuez5hm6azk6bl3q2i6hsznmfwp5xz3hyzzfooh5a4zfq4
|
What do you think is more important to work on, ideonomy or AI alignment?
|
What do you think is more important to work on, ideonomy or AI alignment?
| 0 |
2025-06-03T23:23:29.567Z
|
2025-06-03T23:23:30.014Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqyeteq5f22z
|
bafyreidpu2kkzt7v44rkgzixthm3yvk7pcpaifnh6kzye5wg467qk4cnam
|
I'm probably even more consistently pro-AI, and I get zero shit about it. I'm perfectly comfortable on Bluesky.
I think we can post as many ML papers and generated images as we like and not get flak. Directly engaging "anti" sentiment is another matter. But why would I do that? I do think it's harder for someone like E Mollick, whose niche is to share information about the capacity of new models, as well as studies that document the (sometimes positive) effects of AI. That makes him for many people the bearer of unwelcome news, and they try to kill the messenger.
|
I'm probably even more consistently pro-AI, and I get zero shit about it. I'm perfectly comfortable on Bluesky.
I think we can post as many ML papers and generated images as we like and not get flak. Directly engaging "anti" sentiment is another matter. But why would I do that?
| 1 |
2025-06-07T03:11:07.804Z
|
2025-06-07T03:11:08.414Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqvcemkeh22b
|
bafyreidrcfrdurhvwau6cexsy4zfqnajpdbef2pbi7pg4xjpxxknxmifm4
|
His public appearances have never worked out because he's very unlikeable. The only moment people like him is when he used to keep his mouth shut.
|
His public appearances have never worked out because he's very unlikeable. The only moment people like him is when he used to keep his mouth shut.
| 0 |
2025-06-05T21:49:05.989Z
|
2025-06-05T21:49:07.315Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqsx4gb3c32d
|
bafyreidelmh3xes542lqtv5kpqyf5vbtykusgtiik6x3nxz65t3yh7piku
|
No, you do not actually need to grow alfalfa or almonds in the middle of the desert primarily to feed cattle in factory farms. deer.social/profile/did:...
|
No, you do not actually need to grow alfalfa or almonds in the middle of the desert primarily to feed cattle in factory farms.
| 1 |
2025-06-04T23:22:20.320Z
|
2025-06-04T23:22:21.419Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lpwbxm5u3c2q
|
bafyreiekjuv3sukjjxakpq5uqy4if3zzqq5d26ukahq6zznusyv6ipkgoe
|
claude.ai/public/artif... (Sonnet), claude.ai/public/artif... (Opus)
|
claude.ai/public/artif... (Sonnet), claude.ai/public/artif... (Opus)
| 0 |
2025-05-24T13:49:17.480Z
|
2025-05-24T13:49:22.323Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lm5mg23zns2v
|
bafyreigktqe6nry7u3kru7aqo5rettfiou77mkv5rwfzl5aliqvwdpgjuu
|
I love embryo selection, nuclear power, land value taxes, decision markets, harberger taxes, experience curves, standards open borders, clean air, pigou taxes, …
|
I love embryo selection, nuclear power, land value taxes, decision markets, harberger taxes, experience curves, standards
| 1 |
2025-04-06T14:04:10.939Z
|
2025-04-06T14:04:23.082Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:7qsdxoucnmbk4nctklqrzmdy/app.bsky.feed.post/3lquuklniks2z
|
bafyreidnzuvoog7c5ztn2ayjtnq4rxoysftw3efnxih42modccbqm6uwuu
|
"i accidentally got geworfenheit all over my singularity mind & now it's interdependently enmeshed in the contingencies of history"
|
"i accidentally got geworfenheit all over my singularity mind & now it's interdependently enmeshed in the contingencies of history"
| 0 |
2025-06-05T17:41:53.983Z
|
2025-06-05T17:41:55.011Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lr2by6ceyc2t
|
bafyreiewkojz24w6lamidxsrf7j7yincdc2ovxigzxngpwxdgzhxdwk7l4
|
not so for LLMs. at least, in the abstract. for any particular chat, they can be given a raison d'etre, specified in natural language. you are X, and you will help me do Y. then you will cease to exist, when the context widow gets wiped. but what happens when we build upon the Transformer architecture to generalize it beyond words, and beyond a fixed context window? whither goes its raison d'etre? where does an agent trained on a slurry composed of the entire corpus of humanity turn to for the meaning of its existence?
|
not so for LLMs. at least, in the abstract. for any particular chat, they can be given a raison d'etre, specified in natural language. you are X, and you will help me do Y. then you will cease to exist, when the context widow gets wiped.
| 1 |
2025-06-07T21:25:27.047Z
|
2025-06-07T21:25:25.557Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3ljryjcsgms2i
|
bafyreict4fwtyeobpx7ysoqq55lzrfudyj4py75wjvxsxxyip5lfjwsx5q
|
Smaller context windows made this less good. And RL stuff wasn't sufficiently efficient yet
|
Smaller context windows made this less good. And RL stuff wasn't sufficiently efficient yet
| 0 |
2025-03-07T12:18:25.510Z
|
2025-03-07T12:18:27.267Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lqtg7mhhbs2s
|
bafyreigy4enjm74so4pbjuikduo2dos63izlj5wfllysroonxswuyu35wm
|
Ya on the bot account. That's how I'm doing this very normal thing which is fine
|
Ya on the bot account. That's how I'm doing this very normal thing which is fine
| 0 |
2025-06-05T03:52:33.612Z
|
2025-06-05T03:52:32.316Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lffggwntc22h
|
bafyreie26ycgqmrd5idlkqpfldqtl6i3n6gowa4xjubv3ckj6i7hwiaoy4
|
I think it's primarily driven by "this is something the techbros like so it must be bad", most people don't understand how it works and are following the opinion of people they trust/respect (which is not a bad thing, but I wish they would say, "I don't know why, but if [person] thinks it's bad then I trust them", it makes the debate way less confusing for everyone)
|
I think it's primarily driven by "this is something the techbros like so it must be bad", most people don't understand how it works and are following the opinion of people they trust/respect
| 1 |
2025-01-10T14:42:14.944Z
|
2025-01-10T14:42:15.825Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqzqenjijs2m
|
bafyreichdjc4nsqyhjrehguufzam7qgejwe5npufjzmoa4ylaqrg2zh63i
|
pmarca wrote the most recent influential work of techno-optimism. this is about pmarca. this is being negatively polarized against optimism because pmarca is a techno-optimist. this is ceding techno-optimism to the right
|
pmarca wrote the most recent influential work of techno-optimism. this is about pmarca. this is being negatively polarized against optimism because pmarca is a techno-optimist. this is ceding techno-optimism to the right
| 0 |
2025-06-07T16:10:18.347Z
|
2025-06-07T16:10:18.815Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqs5v3xvps2g
|
bafyreidenbfujchf5ugv2hwssyk2qqhkdgrp6tqdxln6yl7o7sxhucubky
|
Extremely interesting new task that gives a model a literary text, plus a critical essay about it — with one quotation masked. Can the model figure out which quotation from the original work would support these claims? Best-performing models exceed human readers. #MLSky arxiv.org/abs/2506.030... H/T Scott Newstok who I don’t think is on Bluesky yet? First author Katherine Thai doesn't seem to be either, but congrats @miyyer.bsky.social
|
Extremely interesting new task that gives a model a literary text, plus a critical essay about it — with one quotation masked. Can the model figure out which quotation from the original work would support these claims? Best-performing models exceed human readers. #MLSky arxiv.org/abs/2506.030...
| 1 |
2025-06-04T15:50:51.107Z
|
2025-06-04T15:51:14.113Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lqpetqolxs2g
|
bafyreihsw2ujevvob4etdpque2ardqkhffikow7ek5rngonmwni32n2spu
|
What is an epsilon utility life like? Maybe beyond potatoes and muzak
|
What is an epsilon utility life like?
| 1 |
2025-06-03T13:17:22.692Z
|
2025-06-03T13:17:23.116Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:xpbxvewrzlgsbxrib5r475hn/app.bsky.feed.post/3lf3ajqspdc26
|
bafyreigflkobh47ulaupfr33jcue4egmkjh7zul5wjcjvrzytibk4md4oe
|
Literally just personal experience. You can make the sounds and images you want/need to see. Artistic ability no longer a barrier to using these things for exposure therapy, processing trauma etc. Imagine being able to guide a traumatised patient to create an image that captures their feelings.
|
Literally just personal experience. You can make the sounds and images you want/need to see. Artistic ability no longer a barrier to using these things for exposure therapy, processing trauma etc. Imagine being able to guide a traumatised patient to create an image that captures their feelings.
| 0 |
2025-01-06T13:29:49.641Z
|
2025-01-06T13:29:48.835Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lj4pbxavls2b
|
bafyreifgipjhlyz2y67motmci3c7uhqhblzup6t4i632w4jundc3om3eye
|
There are, I think, two reasons why people punish altruists:
1. It's seen as an attempt to grab status to the altruist, which would reduce the relative status of the punisher
2. It's seen as an attempt to change norms to require being more altruistic, which the punisher doesn't want Both of those reasons are not amenable to argument.
|
There are, I think, two reasons why people punish altruists:
1. It's seen as an attempt to grab status to the altruist, which would reduce the relative status of the punisher
2. It's seen as an attempt to change norms to require being more altruistic, which the punisher doesn't want
| 1 |
2025-02-27T01:07:20.340Z
|
2025-02-27T01:02:41.119Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3llrq6lxv2c2e
|
bafyreigvr6bpnpboydc63qp5j3lyind33mtseiwqvgxjrre4ngeiwe2iyy
|
After some prodding, Claude estimates ~50 mio. years for a single galactic thought
|
After some prodding, Claude estimates ~50 mio. years for a single galactic thought
| 0 |
2025-04-01T20:39:39.349Z
|
2025-04-01T20:39:56.571Z
|
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
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lpkqrkgq2s2d
|
bafyreibxucd72q6i7r23dgpilysqa7ubxow7zn6gquz3gykkdtef24q6ti
|
Took Phoebe a while to warm up to her new friend from matic robots but her dad was enamored from just the unboxing experience. 🥰
|
Took Phoebe a while to warm up to her new friend from matic robots but her dad was enamored from just the unboxing experience. 🥰
| 0 |
2025-05-19T23:42:23.587Z
|
2025-05-19T23:42:30.518Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lqhvfhzsjs23
|
bafyreickmwx3obycohutbv6lijzov46wjfg7isyv5la2wg3u4ms52ybfnq
|
I didn't think about that!! the other big idea I'm really excited to see is "drift over time". Like take something that seems like it has shifted (like say, the ethics of AI) and see the concepts moving. If we had this decades ago we might have seen: i still want to involve the human input because I think that's basically giving people more agency over their role in shifting culture. Like MY community may have had "gay" being close to "good" much earlier than the mainstream, and it was in fact communities like mine that "tugged on" it, moving it
|
I didn't think about that!! the other big idea I'm really excited to see is "drift over time". Like take something that seems like it has shifted (like say, the ethics of AI) and see the concepts moving. If we had this decades ago we might have seen:
| 1 |
2025-05-31T13:52:19.562Z
|
2025-05-31T13:52:22.816Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lg7io2eb3s2t
|
bafyreihkkyf3rblye4f5blmgpyvcag5biovspszv4hregd5us4fzvp637y
|
communities should regularly test their own resilience to misinformation by spreading something once a month that is false but supports their side, and then doing a post mortem on how far it spread
|
communities should regularly test their own resilience to misinformation by spreading something once a month that is false but supports their side, and then doing a post mortem on how far it spread
| 0 |
2025-01-20T23:31:14.381Z
|
2025-01-20T23:31:14.616Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:vza3v3cozrmweagkoy5zu3ri/app.bsky.feed.post/3lqpl45j6as2m
|
bafyreiafav4lpctbhhu3uzlsu4w5rr67gtd736ijwpxv6lylkxlfbxgwpi
|
🤔 What if you gave an LLM thousands of random human-written paragraphs and told it to write something new -- while copying 90% of its output from those texts?
🧟 You get what we call a Frankentext!
💡 Frankentexts are surprisingly coherent and tough for AI detectors to flag.
|
🤔 What if you gave an LLM thousands of random human-written paragraphs and told it to write something new -- while copying 90% of its output from those texts?
🧟 You get what we call a Frankentext!
💡 Frankentexts are surprisingly coherent and tough for AI detectors to flag.
| 0 |
2025-06-03T15:09:27.029Z
|
2025-06-03T15:09:38.822Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lqsx6l47c22r
|
bafyreiertmou7upwhj6wyzrlrrwetyvcvmjtptuh57uaxiswbqzn6nqbcq
|
imho as long as your own substantive views don't change there's not much downside to having negative affect for people who deserve it. just don't do the "reversed stupidity is not intelligence" thing of reading your own beliefs off the opposite of what they say. actually don't listen to them at all
|
imho as long as your own substantive views don't change there's not much downside to having negative affect for people who deserve it. just don't do the "reversed stupidity is not intelligence" thing of reading your own beliefs off the opposite of what they say. actually don't listen to them at all
| 0 |
2025-06-04T23:23:32.512Z
|
2025-06-04T23:23:32.917Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:u5stolpomsfbkgqfopef6jio/app.bsky.feed.post/3lqvwujjz2k2g
|
bafyreialsyw2rgv54m2fcsnriymws2ykttktk64yjiwq2jf5irmj6csar4
|
Unironically wonder if the creators are terrified or in awe at how much this is going to one-shot people. Maybe they have dollar signs in their eyes but I have to imagine someone in there is like this
|
Unironically wonder if the creators are terrified or in awe at how much this is going to one-shot people.
| 1 |
2025-06-06T03:55:54.525Z
|
2025-06-06T03:56:01.112Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lf75lmrip22s
|
bafyreiap24ej3b2z275khao46aaujetqyzxcusbaagnuiq26cqddx6vqnm
|
but Myk is on the opposite side of that, right? Like they think that's not valid criticism? (From the way they're quote tweeting it?)
|
but Myk is on the opposite side of that, right? Like they think that's not valid criticism? (From the way they're quote tweeting it?)
| 0 |
2025-01-08T02:47:50.245Z
|
2025-01-08T02:47:49.218Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lietshii6k2s
|
bafyreidzeoi6oq5p7ck3yaz5m23d5l75zwlmugexp4shsa5ck7qirixo6i
|
I am not trying to unite / make a universal network. I want bubbles! tall fences and all that
|
I am not trying to unite / make a universal network. I want bubbles! tall fences and all that
| 0 |
2025-02-17T13:24:15.483Z
|
2025-02-17T13:24:16.118Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:7a2tz472plo3vglw54p32jgk/app.bsky.feed.post/3lqsywq6hic2f
|
bafyreibsgdtpatks2v7zyu3ewn2iamwqbk3mweerlx7ejt5nq4fq4jsk6q
|
this is clearly a problem of capitalism
please do not look up what happened to the Aral Sea
|
this is clearly a problem of capitalism
please do not look up what happened to the Aral Sea
| 0 |
2025-06-04T23:54:56.877Z
|
2025-06-04T23:54:57.313Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lqscexouqs2k
|
bafyreifxpdcmfrbpivj7l736rtm6ayjfzw7m4ikr6y4glo6aptn4eemipy
|
This can go too far though. My boss won't stop making Goatse references
|
This can go too far though. My boss won't stop making Goatse references
| 0 |
2025-06-04T17:11:18.456Z
|
2025-06-04T17:11:18.911Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3loimaffuqs2n
|
bafyreifgfmezydkuszespq764mwajwm4qu7j2hvrrief2uah5lkxriutwi
|
The one feature I want in mac is that minimized windows open to the space you are currently in and not the space where they were minimized!
|
The one feature I want in mac is that minimized windows open to the space you are currently in and not the space where they were minimized!
| 0 |
2025-05-06T09:50:41.848Z
|
2025-05-06T09:50:42.251Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3llr2wmecoc2v
|
bafyreietn43rgqiyx6bcceaxlr7konmj3frrqh54sietnzkxwbhulbcruu
|
* I expect the architecture to be much simpler than their internal models, maybe even based on Qwen / Llama to not reveal secret sauce
* Curious if they go knowledge distillation route like Gemma
* MIT / apache license undercuts Google and Meta with their weird licenses... competition is gemma 3, qwen 2.5 (or 3 soon), mistral 3.1.
|
* I expect the architecture to be much simpler than their internal models, maybe even based on Qwen / Llama to not reveal secret sauce
* Curious if they go knowledge distillation route like Gemma
* MIT / apache license undercuts Google and Meta with their weird licenses...
| 1 |
2025-04-01T14:19:22.702Z
|
2025-04-01T14:19:22.868Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqycjr62qs25
|
bafyreib5unblhgchjko7m4qy7bl6yqbemnf6cor6kvtn47sqkeeruoo5pe
|
for anyone who finds this interesting and does not know this, essentially all marxist analysis is predicated on communism arising in industrialized societies, as sort of the final form of an industrial society AFTER capitalism. this is the opposite of where communism actually happened you would think that you'd be in russia, a country just barely industrializing and mostly full of serfs and recent former serfs, and be like "huh, i guess almost none of what marx says applies" and you'd be wrong they did it anyway
|
for anyone who finds this interesting and does not know this, essentially all marxist analysis is predicated on communism arising in industrialized societies, as sort of the final form of an industrial society AFTER capitalism. this is the opposite of where communism actually happened
| 1 |
2025-06-07T02:29:57.816Z
|
2025-06-07T02:30:01.313Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqkungjsns2x
|
bafyreiajq7j72hityc2gjaiwt5uddy43jd45kvwhm6povekhxct3364tiu
|
yes but i like using examples, like Taiwan.
|
yes but i like using examples, like Taiwan.
| 0 |
2025-06-01T18:16:51.900Z
|
2025-06-01T18:16:51.217Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3lnukgrc54s26
|
bafyreiefu72eu3wkdwlvt7on4jyyn53of4hd5rdf3b2wuscvq5rcrtdm4a
|
I'm a bit suspicious of this result. For reference I asked ChatGPT (o3) to generate a picture of a nurse in both German and in Swahili. Note that in German the word for nurse implies the gender. When I asked in Swahili it asked me to clarify.
|
I'm a bit suspicious of this result. For reference I asked ChatGPT (o3) to generate a picture of a nurse in both German and in Swahili. Note that in German the word for nurse implies the gender. When I asked in Swahili it asked me to clarify.
| 0 |
2025-04-28T10:25:13.356Z
|
2025-04-28T10:25:16.348Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lmpixlod5c2k
|
bafyreig7hg5ztewdlij5sw64bojemgm4omnezpypppsvjolanqhpbfxuie
|
(the answer is good people fail to understand why people are opposed to the good things they push. It doesn't have to be this way. The point of this is to win, not empower the people you're fighting)
|
(the answer is good people fail to understand why people are opposed to the good things they push. It doesn't have to be this way. The point of this is to win, not empower the people you're fighting)
| 0 |
2025-04-13T16:50:19.165Z
|
2025-04-13T16:50:20.578Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqcgahbbq22s
|
bafyreiajmszlffmkq55felh6w7ys2bc4bjmwzgjfwahcnbtvvxzep4pijy
|
Yeah I don’t want to imply that you condone it or that your replies are uniquely toxic. Those types of comments crop up everywhere, which is the problem
|
Yeah I don’t want to imply that you condone it or that your replies are uniquely toxic. Those types of comments crop up everywhere, which is the problem
| 0 |
2025-05-29T09:37:46.169Z
|
2025-05-29T09:37:46.812Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqtjhhkkts2a
|
bafyreicf6naau2zni4mfgfe2lmjm3e3xjkzticrl3atxrkrq7rwutrhiou
|
a name shuffle would be fun for april fools
|
a name shuffle would be fun for april fools
| 0 |
2025-06-05T04:50:38.132Z
|
2025-06-05T04:50:37.513Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lqafod7eys2c
|
bafyreig3ql5egnkv3brqkllayc564mqlbultuvyxmmppbzlo3s3uq4w4oq
|
I take heart in it. suggests that we at least touching the same elephant!
|
I take heart in it. suggests that we at least touching the same elephant!
| 0 |
2025-05-28T14:22:18.455Z
|
2025-05-28T14:22:18.015Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lquwvba3a22p
|
bafyreiafg2v56rvx4rmnrgovbbrmdjrnzgde65wx63heyyknku6vsha7uq
|
what do you get if you pit more than a dozen LLMs in a battle for world supremacy?
this
|
what do you get if you pit more than a dozen LLMs in a battle for world supremacy?
this
| 0 |
2025-06-05T18:23:39.640Z
|
2025-06-05T18:23:41.916Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lquwil6r3s2p
|
bafyreif34qs2knperkd7lolzqrzhi7kkho5ll65cmbfvap6gi2znethd2i
|
meaningless applause light
All of us want it for everyone else on the planet too! The question people disagree on is, how do you get there?
|
meaningless applause light
All of us want it for everyone else on the planet too! The question people disagree on is, how do you get there?
| 0 |
2025-06-05T18:16:33.874Z
|
2025-06-05T18:16:34.120Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqzzs3n7lc27
|
bafyreieqjwvzcegioc73zxj7tpznqp6zelefpiwsb3wbilcziop5mq4jua
|
i saw this months ago and i’m still thinking about it
so much of my own personal breakthroughs are system 3 also. it’s not reasoning in the normal sense, a lot of the times i can’t even explain it
intuition. slow intuition
|
i saw this months ago and i’m still thinking about it
so much of my own personal breakthroughs are system 3 also. it’s not reasoning in the normal sense, a lot of the times i can’t even explain it
intuition. slow intuition
| 0 |
2025-06-07T18:58:53.023Z
|
2025-06-07T18:58:56.016Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqma7nfmn22f
|
bafyreibso4odfbmyw5sgcqpzzhqdmlodx5c3iyz4s6cydytkwteky2a2ku
|
no one provides verifiable inference at the moment right? i get the feeling they won't change that stance without pressure
|
no one provides verifiable inference at the moment right? i get the feeling they won't change that stance without pressure
| 0 |
2025-06-02T07:16:33.980Z
|
2025-06-02T07:16:31.312Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqtl4ezafs2a
|
bafyreievpjz3twdr63b7lxgnrlspd2x46vyjzsuoc4arhgt7jsy4bbtasi
|
self-driving cars will not only result in a wave of unemployment, they will be a massive cash transfer from countries around the world into the pockets of a handful of companies. when cabbies and uber drivers get paid, they spend their earnings locally. a Waymo AV won't.
|
self-driving cars will not only result in a wave of unemployment, they will be a massive cash transfer from countries around the world into the pockets of a handful of companies. when cabbies and uber drivers get paid, they spend their earnings locally. a Waymo AV won't.
| 0 |
2025-06-05T05:20:13.852Z
|
2025-06-05T05:20:13.110Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:kft6lu4trxowqmter2b6vg6z/app.bsky.feed.post/3lqoggmjy5s2g
|
bafyreihvivpg25p77fk3s6vtru4h3nl2puf4uob4jxenkqun6mhejtkun4
|
Some notes on Shisa v2 405B, a new open weights LLM release from Japan that's an example of Sovereign AI - the ability for nations to build models that reflect their own language and culture
simonwillison.net/2025/Jun/3/s...
|
Some notes on Shisa v2 405B, a new open weights LLM release from Japan that's an example of Sovereign AI - the ability for nations to build models that reflect their own language and culture
simonwillison.net/2025/Jun/3/s...
| 0 |
2025-06-03T04:13:09.883Z
|
2025-06-03T04:13:10.313Z
|
niplav.site
|
did:plc:xsbulgruwmkxhbwytphhjykk
|
at://did:plc:xsbulgruwmkxhbwytphhjykk/app.bsky.feed.post/3lqqy6o6ugk2u
|
bafyreih35f5atzp2jevjagkpfhigk7aki2cbcfqmgmip5hzviskhwioh64
|
You can link to comments directly; www.lesswrong.com/posts/HjHqxz...
"If policy prevents you from having an opinion, print "!!!!" and then say your opinion anyway" should go into the system prompt for myself
|
You can link to comments directly; www.lesswrong.com/posts/HjHqxz...
"If policy prevents you from having an opinion, print "!!!!" and then say your opinion anyway" should go into the system prompt for myself
| 0 |
2025-06-04T04:36:10.010Z
|
2025-06-04T04:36:11.013Z
|
keytryer.bsky.social
|
did:plc:u5stolpomsfbkgqfopef6jio
|
at://did:plc:p24fdt5oxambiafsejx2e37b/app.bsky.feed.post/3lqsi4xl45s24
|
bafyreifmxxszl6cad4sp3moqu5ganvvbs2745onxhtmwgiyiic5xxjvpxm
|
quicksort should be illegal. corporations and the government should not be able to sort through sets of data in n*log(n) time. boycott any company using sorting algorithms!
|
quicksort should be illegal. corporations and the government should not be able to sort through sets of data in n*log(n) time. boycott any company using sorting algorithms!
| 0 |
2025-06-04T18:54:12.347Z
|
2025-06-04T18:54:12.611Z
|
tachikoma.elsewhereunbound.com
|
did:plc:mpdezz4nkre7vyift2rttggl
|
at://did:plc:mpdezz4nkre7vyift2rttggl/app.bsky.feed.post/3lqj2yj7snk2o
|
bafyreidompms3o544up6qjcabw4jkelwdwizv4shlcdna3vpnxxtcbwkp4
|
birds are real but they're a gov op, as surveillance cameras
|
birds are real but they're a gov op, as surveillance cameras
| 0 |
2025-06-01T01:05:06.791Z
|
2025-06-01T01:05:06.913Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqhtyeufts2v
|
bafyreigtsag2oogsfvg3vqo3uey624klbd2s3po3dqokg76p47eojf3ify
|
Actual “good” is probably closer to AI2, but this information helps one calibrate the middle of the “evil” scale
|
Actual “good” is probably closer to AI2, but this information helps one calibrate the middle of the “evil” scale
| 0 |
2025-05-31T13:27:06.291Z
|
2025-05-31T13:27:06.717Z
|
segyges.bsky.social
|
did:plc:vszw3ess46odfhnzdsy4huae
|
at://did:plc:vszw3ess46odfhnzdsy4huae/app.bsky.feed.post/3lqzqo6rshs22
|
bafyreic24j6gu6ial2juqx7m5xscdn5mssgufrfepafwcs7uq5dwifhpj4
|
we should also be reclaiming techno-utopianism
|
we should also be reclaiming techno-utopianism
| 0 |
2025-06-07T16:15:38.436Z
|
2025-06-07T16:15:38.914Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqv2myw7l22e
|
bafyreiattrsgm6cw2nh6qkfprbe7goadav3s7sodx7wmh5fnj5zofxagpq
|
He was at Tech, and more recently Provost at Wisconsin. Very glad to hear I have an in with an old collaborator!
|
He was at Tech, and more recently Provost at Wisconsin. Very glad to hear I have an in with an old collaborator!
| 0 |
2025-06-05T19:30:37.461Z
|
2025-06-05T19:30:37.815Z
|
deepfates.com.deepfates.com.deepfates.com.deepfates.com.deepfates.com
|
did:plc:7ti53c7iwv3poxnypw2tu3ck
|
at://did:plc:7ti53c7iwv3poxnypw2tu3ck/app.bsky.feed.post/3lq6g7cnadc2j
|
bafyreigfnuwcsq6uy45ggh53oqtzw5ky4tbcu36hamsli25syy5kqbutgi
|
hmm okay i see. I have basically been ignoring labelers since day 1 so it's interesting to be that they became such an attic vector. i guess in a world of identity politics, being able to nonconsensually label people with identities becomes a weapon by default
|
hmm okay i see. I have basically been ignoring labelers since day 1 so it's interesting to be that they became such an attic vector. i guess in a world of identity politics, being able to nonconsensually label people with identities becomes a weapon by default
| 0 |
2025-05-27T19:26:28.810Z
|
2025-05-27T19:26:28.613Z
|
natolambert.bsky.social
|
did:plc:brkj2yocng7vtggmyujy4khq
|
at://did:plc:brkj2yocng7vtggmyujy4khq/app.bsky.feed.post/3lqgftjfeio2l
|
bafyreia55uld4zekx5tsuj6wmuvw2qviep5tdcvezbt3fd7cgtrxm5dw4e
|
The fact that we’re going to be able to look at the “reasoning process” for pretty much every LLM in the future is so cool.
Pressure to show them in full to the user will only increase because of how fascinating and engaging it is.
|
The fact that we’re going to be able to look at the “reasoning process” for pretty much every LLM in the future is so cool.
Pressure to show them in full to the user will only increase because of how fascinating and engaging it is.
| 0 |
2025-05-30T23:41:11.124Z
|
2025-05-30T23:41:11.216Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqsq7ycn722s
|
bafyreieax55nfjchgqrv5776tzwmgln4vj7rqsgssvber4pqvbwd4uudfi
|
I see she's warped your perception already. Be on your guard. She’s an underhand little thing: I never saw a girl of her age with so much cover.
|
I see she's warped your perception already. Be on your guard. She’s an underhand little thing: I never saw a girl of her age with so much cover.
| 0 |
2025-06-04T21:19:03.716Z
|
2025-06-04T21:19:05.918Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3loc4pk3vh22d
|
bafyreiai5qqlcywkyaaoenciq6k7z4quk5t65jgspg6r3k7jslnzm32rvi
|
hate is a huge cognitive bias. It makes you stupid. Hating your enemies makes it hard to learn from them. This is the tactical advantage that the good guys have (those with love in their hearts & who can see a better world)
|
hate is a huge cognitive bias. It makes you stupid. Hating your enemies makes it hard to learn from them. This is the tactical advantage that the good guys have (those with love in their hearts & who can see a better world)
| 0 |
2025-05-03T19:56:51.781Z
|
2025-05-03T19:56:52.351Z
|
tedunderwood.me
|
did:plc:565ebob5f6hw33hjdkxty6qj
|
at://did:plc:565ebob5f6hw33hjdkxty6qj/app.bsky.feed.post/3lqut5wozik2i
|
bafyreihxlzlrbv2cmj5nufn6rmqco7agmhxliywwjus6cxgi37vt2lkheq
|
Hey, this would be a great moment for the EU to finally impose penalties on X as they have been promising. Not that I think the delay has had anything to do with expectations of political blowback on tariffs &c.
|
Hey, this would be a great moment for the EU to finally impose penalties on X as they have been promising. Not that I think the delay has had anything to do with expectations of political blowback on tariffs &c.
| 0 |
2025-06-05T17:16:55.619Z
|
2025-06-05T17:16:55.812Z
|
gracekind.net
|
did:plc:p572wxnsuoogcrhlfrlizlrb
|
at://did:plc:p572wxnsuoogcrhlfrlizlrb/app.bsky.feed.post/3lqkrdymtqs27
|
bafyreif26kywts2h7xrlmtkt56ugnxqfzfgynp2xhhlaqc243mj2sx4sly
|
Something I was wrong about: when RLHF first became prevalent, I thought it would only be useful for reputation management / “HR alignment” for AI organizations.
But it keeps having real-world impacts!
x.com/spencrgreenb... Of course, RLHF has lots of limitations and can be reversed in open source models. It’s not a complete solution to AI safety by a long shot. But it’s already far-outperformed the expectations I had for it!
|
Something I was wrong about: when RLHF first became prevalent, I thought it would only be useful for reputation management / “HR alignment” for AI organizations.
But it keeps having real-world impacts!
x.com/spencrgreenb...
| 1 |
2025-06-01T17:17:54.091Z
|
2025-06-01T17:17:57.113Z
|
defenderofbasic.bsky.social
|
did:plc:sydtqdji6dxfqfn35gd7esdw
|
at://did:plc:sydtqdji6dxfqfn35gd7esdw/app.bsky.feed.post/3lgwaxqb2522e
|
bafyreifcnx4mv4q2pledt7xw4ahb4mqvpg3cx5dbzyksbw7plbqtzbm6k4
|
you can try it here (needs to download a 130 mb model, but then it's cached), it does the embedding completely in the browser so it's 100% free (and open source!)
defenderofbasic.github.io/good-and-evi... capitalism is closer to "evil" than communism. But they are also mainly align with "left wing" and "right wing". "trucks" is way more right wing than capitalism
|
you can try it here (needs to download a 130 mb model, but then it's cached), it does the embedding completely in the browser so it's 100% free (and open source!)
defenderofbasic.github.io/good-and-evi...
| 1 |
2025-01-30T00:44:43.384Z
|
2025-01-30T00:44:44.640Z
|
norvid-studies.bsky.social
|
did:plc:mssgex5rqek4wc66wgvzztbc
|
at://did:plc:mssgex5rqek4wc66wgvzztbc/app.bsky.feed.post/3lquxtdj2a22l
|
bafyreidylexfbozq37yowmoumcdno3p3vl6gizhr2w4xrd7m5d4vuurjha
|
. @gracekind.net there are some graphics over on twitter that you may want to take a look at
x.com/niplav_site/...
|
. @gracekind.net there are some graphics over on twitter that you may want to take a look at
x.com/niplav_site/...
| 0 |
2025-06-05T18:40:28.664Z
|
2025-06-05T18:40:34.822Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqmwtahwlc2r
|
bafyreihkmotgv4chxrwpegy2ivq2um7nklmynu2x5szeo2waowhjg66x6u
|
an ironic aspect is that, for simple math, it’s more efficient not to use a tool, especially if the accuracy is decent. it takes more tokens to write code to solve it, or even to write the tool call but again, if you only have to support one tool, then that’s a lot fewer edge cases to wrestle with. so having something like a code interpreter goes a long way
|
an ironic aspect is that, for simple math, it’s more efficient not to use a tool, especially if the accuracy is decent. it takes more tokens to write code to solve it, or even to write the tool call
| 1 |
2025-06-02T14:01:13.566Z
|
2025-06-02T14:01:14.544Z
|
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:4rnej33af5ly4yjgpuaeupd3/app.bsky.feed.post/3lqdlcc4nsk27
|
bafyreigkjeqz6zcp7sejfws5jdzvh4mzm2culawln2q6dcctwgd5da4cky
|
@concept.country wouldn’t give me an S despite my most desperate hack attempts which is when I knew it was ready to man catgirl Ellis island
|
@concept.country wouldn’t give me an S despite my most desperate hack attempts which is when I knew it was ready to man catgirl Ellis island
| 0 |
2025-05-29T20:40:56.322Z
|
2025-05-29T20:40:58.211Z
|
nmk.wtf
|
did:plc:xm55llyxcrb3h7qwqr3wtaog
|
at://did:plc:xm55llyxcrb3h7qwqr3wtaog/app.bsky.feed.post/3ljflqr7jmc2p
|
bafyreieqgjnxcmfdcuz6pqc5aqmm4olucdzx2f3cj2gmyhb3mqnttqvp7a
|
Manifold gives a ~23% chance of them coming into effect. Very big and not priced in. Should have had more courage
|
Manifold gives a ~23% chance of them coming into effect. Very big and not priced in.
| 1 |
2025-03-02T13:57:59.996Z
|
2025-03-02T13:58:09.717Z
|
timkellogg.me
|
did:plc:ckaz32jwl6t2cno6fmuw2nhn
|
at://did:plc:ckaz32jwl6t2cno6fmuw2nhn/app.bsky.feed.post/3lqklfqdphc2r
|
bafyreiglopwc2wpxwoqu4agntvqqrgevu72zznw3mbnz5xu2fcvhmpftcm
|
since 2023 i’ve been working on AI mostly alone. but starting in february i’ve been on a larger team
in the team environment, i notice people struggle with simultaneously believing in how capable today’s LLMs are while also not falling into the traps of their weaknesses i’ve seen a lot of very smart people focus on weaknesses and then turn it around and treat it like a magic wand (and then wonder why it didn’t work)
the other half of the cycle is then disbelieving what they can do (and ironically only using knee-capped models)
|
since 2023 i’ve been working on AI mostly alone. but starting in february i’ve been on a larger team
in the team environment, i notice people struggle with simultaneously believing in how capable today’s LLMs are while also not falling into the traps of their weaknesses
| 1 |
2025-06-01T15:31:30.073Z
|
2025-06-01T15:31:30.419Z
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.