| --- Log opened Tue Feb 01 00:00:55 2022 |
| 00:23 | | SmithKurosaki [uid215460@Nightstar-e2nmdb.irccloud.com] has joined #code |
| 01:12 | < Kizor_> | "Creatures bred for speed grow really tall and generate high velocities by falling over" |
| 01:14 | < Kizor_> | "RL agent that is allowed to modify its own body learns to have extremely long legs that allow it to fall forward and reach the goal." |
| 01:14 | < Kizor_> | "Genetic algorithm is supposed to configure a circuit into an oscillator, but instead makes a radio to pick up signals from neighboring computers" |
| 01:15 | <&McMartin> | Oh man that last one |
| 01:16 | <&McMartin> | AIUI using genetic algorithms (or, uh, machine learning, as we call it these days) to create circuits tends to produce *absolutely rampant* exploitation of parasitic capacitance and such in lieu of actually having intentional capacitors |
| 01:16 | < Alek> | technically, being tall with long legs worked out pretty good for humans, no? |
| 01:16 | <&McMartin> | We don't fall forward as often though |
| 01:16 | < Alek> | that's literally what walking and running is. a controlled forward fall. |
| 01:17 | <&McMartin> | I'm reading Kizor's description as "be as tall as the course, and then instead of running the course, fall on your face" |
| 01:17 | <&McMartin> | I would however like to note that in Bethesda RPGs, speedrunners take the tallest classes because they move faster |
| 01:17 | <&McMartin> | IIRC, moving faster so that perceived motion speed of the ground is the same for all classes |
| 01:18 | < Alek> | hah |
| 01:18 | < Alek> | good to know, thanks |
| 01:18 | <&McMartin> | s/classes/races/ sorry. Classes don't alter height. |
| 01:20 | < Kizor_> | "AI trained to classify skin lesions as potentially cancerous learns that lesions photographed next to a ruler are more likely to be malignant." |
| 01:21 | < Alek> | ha |
| 01:26 | <&McMartin> | What are you consulting here? |
| 01:27 | | * McMartin also needs to relink the experiment that gave us "Clardic Fug". https://www.aiweirdness.com/new-paint-colors-invented-by-neural-17-05-17/ |
| 01:28 | | gizmore|2 [kvirc@Nightstar-4hh00p.dip0.t-ipconnect.de] has joined #code |
| 01:32 | | gizmore [kvirc@Nightstar-fdlhrv.dip0.t-ipconnect.de] has quit [Ping timeout: 121 seconds] |
| 02:01 | < Kizor_> | McMartin: https://docs.google.com/spreadsheets/u/1/d/e/2PACX-1vRPiprOaC3HsCf5Tuum8bRfzYUiKLRqJmbOoC-32JorNdfyTiRRsR7Ea5eWtvsWzuxo8bjOxCG84dAg/pubhtml |
| 02:46 | <&McMartin> | Thanks |
| 04:10 | | Degi [Degi@Nightstar-41k9gn.pool.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 04:13 | | Degi [Degi@Nightstar-j01ouu.pool.telefonica.de] has joined #code |
| 04:25 | | Kizor_ [a@Nightstar-nfsqa7.yok.fi] has quit [Ping timeout: 121 seconds] |
| 05:14 | | mac [macdjord@Nightstar-re5.7if.45.45.IP] has quit [[NS] Quit: Deep inside, every housecat believes themself to be just a temporarily embarrassed tiger.] |
| 05:18 | | macdjord [macdjord@Nightstar-re5.7if.45.45.IP] has joined #code |
| 05:18 | | mode/#code [+o macdjord] by ChanServ |
| 06:05 | | Vorntastic [uid293981@Nightstar-phvupn.irccloud.com] has joined #code |
| 06:05 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 06:39 | | SmithKurosaki [uid215460@Nightstar-e2nmdb.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 07:02 | | Kindamoody[zZz] is now known as Kindamoody|afk |
| 07:05 | | Pinkhair [Pink@Nightstar-oetkb4.ph.cox.net] has quit [Ping timeout: 121 seconds] |
| 09:21 | | Kizor [a@Nightstar-nfsqa7.yok.fi] has joined #code |
| 09:55 | | Alek [Alek@Nightstar-06ca3p.il.comcast.net] has quit [[NS] Quit: ] |
| 09:56 | | Alek [Alek@Nightstar-06ca3p.il.comcast.net] has joined #code |
| 10:14 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 10:14 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 10:21 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [[NS] Quit: -a- IRC for Android 2.1.59] |
| 10:25 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 12:38 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 12:38 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 13:23 | <@gnolam> | My favorite "machine learning in medicine" tale is the so-called-AI that got insanely good (better than human radiologists) at recognizing liver (?) cancer in scans. |
| 13:24 | <@gnolam> | Only when they actually dissected its decision process they realized that it wasn't in fact looking at the scans /at all/ - it had access to the full metadata and not just the image itself. |
| 13:24 | <@gnolam> | And so, as that information was available in both the training data and the test data, its decision process essentially boiled down to "IF patient referred to specialist cancer clinic THEN cancer ELSE not cancer". |
| 13:47 | <~Vorntastic> | Lol |
| 13:59 | < catalyst> | heh heh |
| 14:07 | | Netsplit Deepthought.Nightstar.Net <-> Krikkit.Nightstar.Net quits: @Syloq, @PinkFreud |
| 14:11 | | Netsplit over, joins: Syloq |
| 14:17 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 14:17 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection reset by peer] |
| 14:22 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 14:22 | | ServerMode/#code [+o PinkFreud] by *.Nightstar.Net |
| 14:22 | | mode/#code [+o Syloq] by ChanServ |
| 14:51 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 14:52 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:40 | <@Tamber> | Yeah, that's the 'great' thing about machine learning. It might 'learn' something, but are you *really* sure it's learned what you *think* it's learned? |
| 16:04 | | Kindamoody|afk is now known as Kindamoody |
| 16:05 | | Vorntastic [uid293981@Nightstar-phvupn.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 16:59 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
| 17:24 | | abudhabi__ [abudhabi@Nightstar-5c96j9.adsl.tpnet.pl] has joined #code |
| 17:24 | < abudhabi__> | Man, the Linux display stuff is just bad. |
| 17:24 | < abudhabi__> | I've been grappling with resolutions and everywhere there's some corner case that breaks things. |
| 17:25 | < abudhabi__> | My monitor malfunctioned, so I connected up the shittier spare one. |
| 17:26 | < abudhabi__> | It also supports full HD, but on Linux it gives visual artifacts (a few flickering, short, one-pixel lines near the bottom) but on Windows it works fine. |
| 17:26 | < abudhabi__> | So I'm resorting to a shitty non-native resolution while I wait for my new monitor to arrive. |
| 17:33 | < abudhabi__> | The biggest gripe with things as they are I have is that resolutions that are supported are nonetheless not detected. |
| 17:34 | < abudhabi__> | I'm not sure how that works, but Windows somehow manages this feat. |
| 17:52 | <&[R]> | Have you done a memtest? |
| 18:11 | < abudhabi__> | No. |
| 18:14 | | * [R] suggests doing so, weird stuff like that screams memory issues |
| 19:02 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 19:04 | | catalyst [catalyst@Nightstar-o5mj7a.dab.02.net] has joined #code |
| 19:07 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Ping timeout: 121 seconds] |
| 19:16 | | catalyst_ [catalyst@Nightstar-058ijn.cable.virginm.net] has joined #code |
| 19:19 | | catalyst [catalyst@Nightstar-o5mj7a.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 22:16 | | catalyst [catalyst@Nightstar-o5mj7a.dab.02.net] has joined #code |
| 22:19 | | catalyst_ [catalyst@Nightstar-058ijn.cable.virginm.net] has quit [Ping timeout: 121 seconds] |
| 22:25 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 22:25 | <&[R]> | https://en.wikipedia.org/wiki/UTF-8#MySQL_utf8mb3 |
| 22:25 | <&[R]> | I found that out today |
| 22:28 | < Mahal> | I'd say driver issues in fact |
| 22:28 | | catalyst [catalyst@Nightstar-o5mj7a.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 22:49 | <&McMartin> | I see that article also refers to WTF-8, though it doesn't give WTF-8 quite enough credit |
| 22:50 | <&McMartin> | What they describe is closer to CESU-8. WTF-8 will correctly handle string concatenation when a UTF-16 code pair has been split mid-character. |
| 23:34 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
| --- Log closed Wed Feb 02 00:00:57 2022 |