3 ms·

How much would it cost to fine-tune an AI to recognize fish and do this automatically instead of crowdsourcing? Maybe some one-shot object detection model woul

How much would it cost to fine-tune an AI to recognize fish and do this automatically instead of crowdsourcing? Maybe some one-shot object detection model would work even without fine-tuning?

__alexs 2y ago
This is about teaching people about waterways and local wild life. Hiding it behind the an AI would defeat the point.
schoen 2y ago
I bet it was a conscious decision to crowdsource it instead of automating it in order to get people more interested in the fish. It seems like it's drawing a pretty substantial fan base!
hhh 2y ago
probably low, but then it would be boring instead of a community event full of whimsy and fun
vasco 2y ago
This is more about community engagement and education than anything else.
noosphr 2y ago
>How much would it cost to fine-tune an AI to recognize fish and do this automatically instead of crowdsourcing? $0 basically. You can fine tune a YOLO model on a work station in an hour or so.
dmbche 2y ago
"People do so enthusiastically every year. In 2024, the Fish Doorbell attracted around 2.7 million viewers, from America to Brazil! The project has gained global recognition, helping people worldwide learn about fish migration and Utrecht’s underwater world." I don't think solving the problem is the goal, it's attracting attention while solving the problem, which is a different goal
brian-armstrong 2y ago
Eager to abandon your observation duties already? But you just got here.
- cluckindan 2y ago
  I like watching cats more than cat food.
henrikschroder 2y ago
*sigh* "I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do laundry and dishes." This thing is whimsy and fun, it's not a chore to be automated away.
amelius 2y ago
This would come into existence almost instantly if there was money to be made with pressing the doorbell at the right moment.
talkingtab 2y ago
Some of the very best moments of my life have come from those moments when I feel like I a part of something bigger than myself. Act in a play. Be the 30th person in a start up. Sing in a choir. Ring a bell. It appears to me that far from enhancing these moments, in general the internet has inhibited them. The focus of our current internet is Me. Instagram where you can show people how beautiful you are. Facebook, where the important thing is what you like. And who you follow. Me. Me. Me. So it seems natural that the "next big thing" is something that will make the internet even less hospitable for a sense of community - for the We. No doorbell, just an AI. But maybe that is a good thing? This internet with all the tracking and lack of civility can become just a place where AI's can spend hours tracking and targeting other AI's. Far out! We can build something else.
- morphle 2y ago
  "It is the destiny of computers to become interactive intellectual amplifiers for all people pervasively networked worldwide." JCR Licklider initially called it the intergalactic network, now know as the internet. >We can build something else Some of us tried (with internet fish swimming around) : https://www.youtube.com/watch?v=cXGLOiZUZ2U https://www.youtube.com/watch?v=cXGLOiZUZ2U https://www.youtube.com/watch?v=1s9ldlqhVkM https://www.youtube.com/watch?v=1s9ldlqhVkM https://tinlizzie.org/VPRIPapers/tr2003001_croq_collab.pdf https://tinlizzie.org/VPRIPapers/tr2003001_croq_collab.pdf
  - talkingtab 2y ago
    Far Out! Thanks. We could use some Complex Adaptive System architecture to create galaxies. Or we could call them Gestalts instead of CAS. In gestalts entities, use rules to interact with a read/write message bus. Ants. Pheromone trails. Write when you find food, not if you don't. We have the pieces for this. Entities - people. We have a r/w bus - the internet. What we don't have are rules. We can think of rules - the kind Wolfram talks about in NKofS - instead as a language. A language in the sense of Sapir-Whorf. "Twitter and Tear Gas" (Zeynep Tufecki) is a good illustration of Sapir-Whorf in internet languages. (Available as a pdf I think) That language is the missing link. In Real Life there is a language for collaboration. Perhaps the task is to understand the elements of IRL collaboration and then transpose it. One could make some guesses about fundamentals. As we have all discovered, one fundamental is trust. IRL Trust is identity and reputation based. With many caveats. Perhaps we could start by considering A New Kind of Identity. ?

How much would it cost to fine-tune an AI to recognize fish and do this automatically instead of crowdsourcing? Maybe some one-shot object detection model would work even without fine-tuning?

__alexs 2y ago

This is about teaching people about waterways and local wild life. Hiding it behind the an AI would defeat the point.

schoen 2y ago

I bet it was a conscious decision to crowdsource it instead of automating it in order to get people more interested in the fish. It seems like it's drawing a pretty substantial fan base!

hhh 2y ago

probably low, but then it would be boring instead of a community event full of whimsy and fun

vasco 2y ago

This is more about community engagement and education than anything else.

noosphr 2y ago

>How much would it cost to fine-tune an AI to recognize fish and do this automatically instead of crowdsourcing? $0 basically. You can fine tune a YOLO model on a work station in an hour or so.

dmbche 2y ago

"People do so enthusiastically every year. In 2024, the Fish Doorbell attracted around 2.7 million viewers, from America to Brazil! The project has gained global recognition, helping people worldwide learn about fish migration and Utrecht’s underwater world." I don't think solving the problem is the goal, it's attracting attention while solving the problem, which is a different goal

brian-armstrong 2y ago

Eager to abandon your observation duties already? But you just got here.

cluckindan 2y ago
I like watching cats more than cat food.

henrikschroder 2y ago

*sigh* "I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do laundry and dishes." This thing is whimsy and fun, it's not a chore to be automated away.

amelius 2y ago

This would come into existence almost instantly if there was money to be made with pressing the doorbell at the right moment.

talkingtab 2y ago

Some of the very best moments of my life have come from those moments when I feel like I a part of something bigger than myself. Act in a play. Be the 30th person in a start up. Sing in a choir. Ring a bell. It appears to me that far from enhancing these moments, in general the internet has inhibited them. The focus of our current internet is Me. Instagram where you can show people how beautiful you are. Facebook, where the important thing is what you like. And who you follow. Me. Me. Me. So it seems natural that the "next big thing" is something that will make the internet even less hospitable for a sense of community - for the We. No doorbell, just an AI. But maybe that is a good thing? This internet with all the tracking and lack of civility can become just a place where AI's can spend hours tracking and targeting other AI's. Far out! We can build something else.

morphle 2y ago
"It is the destiny of computers to become interactive intellectual amplifiers for all people pervasively networked worldwide." JCR Licklider initially called it the intergalactic network, now know as the internet. >We can build something else Some of us tried (with internet fish swimming around) : https://www.youtube.com/watch?v=cXGLOiZUZ2U https://www.youtube.com/watch?v=cXGLOiZUZ2U https://www.youtube.com/watch?v=1s9ldlqhVkM https://www.youtube.com/watch?v=1s9ldlqhVkM https://tinlizzie.org/VPRIPapers/tr2003001_croq_collab.pdf https://tinlizzie.org/VPRIPapers/tr2003001_croq_collab.pdf
- talkingtab 2y ago
  Far Out! Thanks. We could use some Complex Adaptive System architecture to create galaxies. Or we could call them Gestalts instead of CAS. In gestalts entities, use rules to interact with a read/write message bus. Ants. Pheromone trails. Write when you find food, not if you don't. We have the pieces for this. Entities - people. We have a r/w bus - the internet. What we don't have are rules. We can think of rules - the kind Wolfram talks about in NKofS - instead as a language. A language in the sense of Sapir-Whorf. "Twitter and Tear Gas" (Zeynep Tufecki) is a good illustration of Sapir-Whorf in internet languages. (Available as a pdf I think) That language is the missing link. In Real Life there is a language for collaboration. Perhaps the task is to understand the elements of IRL collaboration and then transpose it. One could make some guesses about fundamentals. As we have all discovered, one fundamental is trust. IRL Trust is identity and reputation based. With many caveats. Perhaps we could start by considering A New Kind of Identity. ?