J Λ M Ξ S

If I had to live my life again, I'd make the same mistakes, only sooner.
--
🖖🤘✌️

About Me: https://jmsdnns.com
Code: https://github.com/jmsdnns
Band: https://soundcloud.com/americanf00d

Public Key

npub1sv4r2lnugqz7f98jlwzqx7zz3ugt2hkzpzl2fpc82c7kpfv2rukqgy6mcs

NIP-05 Address

jmsdnns_at_mastodon.social@momostr.pink

Profile Code

nprofile1qqsgx2340e7yqp0yjne0hpqr0ppg7y94tmpq304ysur4v0tq5k9p7tqpr9mhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmnt9uq3samnwvaz7tmjv4kxz7fwd4hk6mmnw3ezuurfde4sscqrg5

Publishing to

relay.momostr.pink

relay.primal.net

Author Public Key

npub1sv4r2lnugqz7f98jlwzqx7zz3ugt2hkzpzl2fpc82c7kpfv2rukqgy6mcs

Show more details

Last Notes

2024-09-28 17:30:17

- reply

classic play.
it's also why i have had so much trouble generating ideas for businesses based on using APIs for AI... like, consider Klarna's claim that openai now does 90% if their business. their entire cost if doing business will go up proportionately

2024-09-26 03:59:39

- reply

that's what I wanna see! privacy + AI

2024-09-22 03:43:30

- reply

fascinating analogy for what Im seeing with AI right now
im long term optimistic about AI, but short term frustrated

2024-09-21 16:44:25

- reply

I enjoy mocking the people who use it. Some examples:
- Programming is too hard for many people, so I get why AI is appealing
- Do you feel comfortable letting them sieze the means of production?
- Wharton researchers have shown AI offers the most benefit to the lowest performers

2024-09-15 16:47:42

💻 The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed
> By some estimates, more than 80 percent of AI projects fail—twice the rate of failure for information technology projects that do not involve AI
> Our interviews highlighted five leading root causes of the failure of AI projects
https://www.rand.org/pubs/research_reports/RRA2680-1.html

2024-09-01 02:07:09

- reply

his videos are phenomenal

2024-09-01 00:40:12

- reply

i am thrilled by what yall are doing! haters gonna hate

2024-08-23 19:36:35

- reply

we're already human shaped! making humanoids is kinda like saying we dont wanna solve new problems, just the old ones

2024-08-23 16:58:35

humanoid robots are so ridiculous. they're the AGI of robotics
COME AT ME

2024-08-22 03:41:18

I love that Tim Walz's son keeps crying
IT'S OK TO FEEL

2024-08-18 18:04:12

- reply

this is fascinating. i have spent significant time considering what genAI means relative to nostalgia. one way I look at genAI now is as a nostalgia generator. whether or not nostalgia is good for society is a fantastic question too

2024-08-18 04:21:11

First time using Lima for virtual machines
Really digging how simple this experience is
https://lima-vm.io/

2024-08-10 21:09:29

🏴‍☠️ Nearly 3 Billion People Hacked in National Public Data Breach.
THREE BILLION!!
https://www.kiplinger.com/personal-finance/billions-hacked-in-national-public-data-breach

2024-08-10 17:23:59

- reply

you just made more data that needs sciencing.

2024-06-16 23:49:18

- reply

excellent post! i like a lot of what I have read on that blog, or heard on their podcast.
here is a tangentially relevant post about a concept from my favorite michael mauboussin book https://fs.blog/the-paradox-of-skill/

2024-06-16 02:53:33

🤖 The Next Generation of Neural Networks, by Geoffrey Hinton
This video is from 2007! The comments are a fascinating time capsule too.
https://youtu.be/AyzOUbkUf3M

2024-06-12 01:44:43

- reply

maybe they heard your sighs
https://files.mastodon.social/media_attachments/files/112/601/116/173/098/242/original/8679041f1eea6142.jpeg

2024-05-23 14:18:45

- reply

fwiw i have been trying searxng while ddg is down and so far it's pretty solid. it's a metasearch and anyone can run an instance. open source too! https://github.com/searxng/searxng

2024-05-23 13:36:26

I know #DuckDuckGo isn't perfect, but I love what they're trying to do so much.
Let's go easy on them for whatever mistakes they make today as they work under extreme stress to get everything working again.
https://files.mastodon.social/media_attachments/files/112/490/668/518/027/726/original/df9335b6ae8f741f.mp4

2024-05-23 13:36:26

2024-05-18 01:13:09

The experiment we just did at Penn to see if AIs can do homeworks for AI students concludes with everyone reassured that AIs completely fail at getting the logic right for anything complex, even with good feedback for improvements
I tried having it generate code for markov decision processes, reinforcement learning, basic language models, and perceptrons, and the AIs did an absolutely terrible job
You can see the homeworks we tested on here http://artificial-intelligence-class.org/