Redstone1080 wrote:

If/when Scratch 4.0 comes out, will there be people doing the whole “3.0 Sucks” and Thanos Cat thing again, but for 4.0?
/hj

depends on how much they change. if they add the ai chatbot thing, there'll certainly be people complaining

Redstone1080 wrote:

If/when Scratch 4.0 comes out, will there be people doing the whole “3.0 Sucks” and Thanos Cat thing again, but for 4.0?
/hj

Yes.

You can't ever make everyone happy- so while it might not be as wide scale as when 3.0 came out, some people will still complain about it.

Redstone1080 wrote:

If/when Scratch 4.0 comes out, will there be people doing the whole “3.0 Sucks” and Thanos Cat thing again, but for 4.0?
/hj

Thanos Cat? What type of cat is this? Never Heard of

If Scratch 4.0 would have Artificial Intelligence in the code (Either ChatGPT, Gemini AI, Copilot, Claude ,etc or just its own AI), Redesigned UI elements, post section in the every scratchers' profile, it would be great!!!

scratch Cloud variables versus Cloudlink for turbo warp

Pushkarajcoder wrote:

Redstone1080 wrote:

If/when Scratch 4.0 comes out, will there be people doing the whole “3.0 Sucks” and Thanos Cat thing again, but for 4.0?
/hj

Thanos Cat? What type of cat is this? Never Heard of

It was a studio made soon after 3.0 was released, and the joke was that Scratch thanos snapped half the users lol

I haven't used Scratch seriously in over a year, so now as an outside perspective, I have to say that adding generative AI to the Scratch platform would absolutely decimate the little soul Scratch has left.

Back in 2020 when I was the most active, Scratch felt like a big world full of imagination and creativity (without the existence of today's AI image generation, I'll add). Every day there was something new on the Explore page, and that creativity and “freshness” honestly got me through the pandemic. Comparatively however, today I see very little of what the platform was in its highest years for me. Introducing AI would just ‘kill’ the site entirely. Nothing would be truly original anymore.

Although, I do believe that other forms of AI, like pathfinding AIs and pattern-finding AIs, would be a much safer addition to Scratch. These types of artificial intelligence don't harm the environment NEARLY as much compared to generative AI. Datacenters use so much water; it directly harms families and homes connected to the same water supply, and most datacenters are intentionally placed in impoverished or rural areas. These same datacenters also contribute highly to the climate change crisis due to the insane amounts of electricity these buildings need to power their machines, and most of this energy comes from oil rigs or coal. In comparison, pathfinding AIs are able to run locally on your personal computer and have already been remade in Scratch many times, so the logical conclusion should be to give Scratchers an easy way to introduce fun, local artificial intelligence to their games… Right?

I don't exactly expect everyone to agree with me on all my expressed opinions here, but I tried my best to include actual facts instead of “what I feel.” Generative AI is already a very controversial topic among many artists, musicians, and programmers. However, when push comes to shove, Scratch adding “next-level AI generation” with the amount of people that use Scratch could result in a high risk of pollution and a drought of real, human creativity on the site.

For clarification, I believe that Scratch is just missing the amount of originality & community it had in 2020

Redstone1080 wrote:

(#140)
If/when Scratch 4.0 comes out, will there be people doing the whole “3.0 Sucks” and Thanos Cat thing again, but for 4.0?
/hj

i guarantee it lol

silvxrcat wrote:

I haven't used Scratch seriously in over a year, so now as an outside perspective, I have to say that adding generative AI to the Scratch platform would absolutely decimate the little soul Scratch has left.

Back in 2020 when I was the most active, Scratch felt like a big world full of imagination and creativity (without the existence of today's AI image generation, I'll add). Every day there was something new on the Explore page, and that creativity and “freshness” honestly got me through the pandemic. Comparatively however, today I see very little of what the platform was in its highest years for me. Introducing AI would just ‘kill’ the site entirely. Nothing would be truly original anymore.

Although, I do believe that other forms of AI, like pathfinding AIs and pattern-finding AIs, would be a much safer addition to Scratch. These types of artificial intelligence don't harm the environment NEARLY as much compared to generative AI. Datacenters use so much water; it directly harms families and homes connected to the same water supply, and most datacenters are intentionally placed in impoverished or rural areas. These same datacenters also contribute highly to the climate change crisis due to the insane amounts of electricity these buildings need to power their machines, and most of this energy comes from oil rigs or coal. In comparison, pathfinding AIs are able to run locally on your personal computer and have already been remade in Scratch many times, so the logical conclusion should be to give Scratchers an easy way to introduce fun, local artificial intelligence to their games… Right?

I don't exactly expect everyone to agree with me on all my expressed opinions here, but I tried my best to include actual facts instead of “what I feel.” Generative AI is already a very controversial topic among many artists, musicians, and programmers. However, when push comes to shove, Scratch adding “next-level AI generation” with the amount of people that use Scratch could result in a high risk of pollution and a drought of real, human creativity on the site.

For clarification, I believe that Scratch is just missing the amount of originality & community it had in 2020

I agree with most of that, generative ai should be really thought over weather it is worth the energy cost, I believe stuff that cannot be easily done in scratch should be added instead.

Here's my suggestion: a pathfinding block that is very inefficient so as to get users to invent more efficient ways to do stuff

silvxrcat wrote:

I haven't used Scratch seriously in over a year, so now as an outside perspective, I have to say that adding generative AI to the Scratch platform would absolutely decimate the little soul Scratch has left.

Back in 2020 when I was the most active, Scratch felt like a big world full of imagination and creativity (without the existence of today's AI image generation, I'll add). Every day there was something new on the Explore page, and that creativity and “freshness” honestly got me through the pandemic. Comparatively however, today I see very little of what the platform was in its highest years for me. Introducing AI would just ‘kill’ the site entirely. Nothing would be truly original anymore.

Although, I do believe that other forms of AI, like pathfinding AIs and pattern-finding AIs, would be a much safer addition to Scratch. These types of artificial intelligence don't harm the environment NEARLY as much compared to generative AI. Datacenters use so much water; it directly harms families and homes connected to the same water supply, and most datacenters are intentionally placed in impoverished or rural areas. These same datacenters also contribute highly to the climate change crisis due to the insane amounts of electricity these buildings need to power their machines, and most of this energy comes from oil rigs or coal. In comparison, pathfinding AIs are able to run locally on your personal computer and have already been remade in Scratch many times, so the logical conclusion should be to give Scratchers an easy way to introduce fun, local artificial intelligence to their games… Right?

I don't exactly expect everyone to agree with me on all my expressed opinions here, but I tried my best to include actual facts instead of “what I feel.” Generative AI is already a very controversial topic among many artists, musicians, and programmers. However, when push comes to shove, Scratch adding “next-level AI generation” with the amount of people that use Scratch could result in a high risk of pollution and a drought of real, human creativity on the site.

For clarification, I believe that Scratch is just missing the amount of originality & community it had in 2020

There're plenty of generative AI models that can be run locally without internet. That's not really the issue most people should have with AI; rather, it's that training the AI is very extremely expensive. A locally-run generative AI would do nothing to address this issue.

I also think that calling pathfinding algorithms “AI” is a huge over-inclusion of the term. For pathfinding, just programming it with regular code is much more efficient than using an AI (accounting for the amount of effort that making an AI takes). A much better example of “AI” would be the Face Sensing extension that's expected to release in the fall. That's been received pretty positively as far as I'm aware, so I don't think that the inclusion of such an “AI” would be a problem to anyone.

ispretty wrote:

silvxrcat wrote:

I haven't used Scratch seriously in over a year, so now as an outside perspective, I have to say that adding generative AI to the Scratch platform would absolutely decimate the little soul Scratch has left.

Back in 2020 when I was the most active, Scratch felt like a big world full of imagination and creativity (without the existence of today's AI image generation, I'll add). Every day there was something new on the Explore page, and that creativity and “freshness” honestly got me through the pandemic. Comparatively however, today I see very little of what the platform was in its highest years for me. Introducing AI would just ‘kill’ the site entirely. Nothing would be truly original anymore.

Although, I do believe that other forms of AI, like pathfinding AIs and pattern-finding AIs, would be a much safer addition to Scratch. These types of artificial intelligence don't harm the environment NEARLY as much compared to generative AI. Datacenters use so much water; it directly harms families and homes connected to the same water supply, and most datacenters are intentionally placed in impoverished or rural areas. These same datacenters also contribute highly to the climate change crisis due to the insane amounts of electricity these buildings need to power their machines, and most of this energy comes from oil rigs or coal. In comparison, pathfinding AIs are able to run locally on your personal computer and have already been remade in Scratch many times, so the logical conclusion should be to give Scratchers an easy way to introduce fun, local artificial intelligence to their games… Right?

I don't exactly expect everyone to agree with me on all my expressed opinions here, but I tried my best to include actual facts instead of “what I feel.” Generative AI is already a very controversial topic among many artists, musicians, and programmers. However, when push comes to shove, Scratch adding “next-level AI generation” with the amount of people that use Scratch could result in a high risk of pollution and a drought of real, human creativity on the site.

For clarification, I believe that Scratch is just missing the amount of originality & community it had in 2020

There're plenty of generative AI models that can be run locally without internet. That's not really the issue most people should have with AI; rather, it's that training the AI is very extremely expensive. A locally-run generative AI would do nothing to address this issue.

I also think that calling pathfinding algorithms “AI” is a huge over-inclusion of the term. For pathfinding, just programming it with regular code is much more efficient than using an AI (accounting for the amount of effort that making an AI takes). A much better example of “AI” would be the Face Sensing extension that's expected to release in the fall. That's been received pretty positively as far as I'm aware, so I don't think that the inclusion of such an “AI” would be a problem to anyone.

pretty sure thats the type of 'AI" they were talking about, stuff that is not generative

wvzack wrote:

ispretty wrote:

silvxrcat wrote:

I haven't used Scratch seriously in over a year, so now as an outside perspective, I have to say that adding generative AI to the Scratch platform would absolutely decimate the little soul Scratch has left.

Back in 2020 when I was the most active, Scratch felt like a big world full of imagination and creativity (without the existence of today's AI image generation, I'll add). Every day there was something new on the Explore page, and that creativity and “freshness” honestly got me through the pandemic. Comparatively however, today I see very little of what the platform was in its highest years for me. Introducing AI would just ‘kill’ the site entirely. Nothing would be truly original anymore.

Although, I do believe that other forms of AI, like pathfinding AIs and pattern-finding AIs, would be a much safer addition to Scratch. These types of artificial intelligence don't harm the environment NEARLY as much compared to generative AI. Datacenters use so much water; it directly harms families and homes connected to the same water supply, and most datacenters are intentionally placed in impoverished or rural areas. These same datacenters also contribute highly to the climate change crisis due to the insane amounts of electricity these buildings need to power their machines, and most of this energy comes from oil rigs or coal. In comparison, pathfinding AIs are able to run locally on your personal computer and have already been remade in Scratch many times, so the logical conclusion should be to give Scratchers an easy way to introduce fun, local artificial intelligence to their games… Right?

I don't exactly expect everyone to agree with me on all my expressed opinions here, but I tried my best to include actual facts instead of “what I feel.” Generative AI is already a very controversial topic among many artists, musicians, and programmers. However, when push comes to shove, Scratch adding “next-level AI generation” with the amount of people that use Scratch could result in a high risk of pollution and a drought of real, human creativity on the site.

For clarification, I believe that Scratch is just missing the amount of originality & community it had in 2020

There're plenty of generative AI models that can be run locally without internet. That's not really the issue most people should have with AI; rather, it's that training the AI is very extremely expensive. A locally-run generative AI would do nothing to address this issue.

I also think that calling pathfinding algorithms “AI” is a huge over-inclusion of the term. For pathfinding, just programming it with regular code is much more efficient than using an AI (accounting for the amount of effort that making an AI takes). A much better example of “AI” would be the Face Sensing extension that's expected to release in the fall. That's been received pretty positively as far as I'm aware, so I don't think that the inclusion of such an “AI” would be a problem to anyone.

pretty sure thats the type of 'AI" they were talking about, stuff that is not generative

Non-generative AI has nothing to do with this topic

ispretty wrote:

wvzack wrote:

ispretty wrote:

silvxrcat wrote:

-snip-

There're plenty of generative AI models that can be run locally without internet. That's not really the issue most people should have with AI; rather, it's that training the AI is very extremely expensive. A locally-run generative AI would do nothing to address this issue.

I also think that calling pathfinding algorithms “AI” is a huge over-inclusion of the term. For pathfinding, just programming it with regular code is much more efficient than using an AI (accounting for the amount of effort that making an AI takes). A much better example of “AI” would be the Face Sensing extension that's expected to release in the fall. That's been received pretty positively as far as I'm aware, so I don't think that the inclusion of such an “AI” would be a problem to anyone.

pretty sure thats the type of 'AI" they were talking about, stuff that is not generative

Non-generative AI has nothing to do with this topic

Why not

ispretty wrote:

silvxrcat wrote:

snip

There're plenty of generative AI models that can be run locally without internet. That's not really the issue most people should have with AI; rather, it's that training the AI is very extremely expensive. A locally-run generative AI would do nothing to address this issue.

I also think that calling pathfinding algorithms “AI” is a huge over-inclusion of the term. For pathfinding, just programming it with regular code is much more efficient than using an AI (accounting for the amount of effort that making an AI takes). A much better example of “AI” would be the Face Sensing extension that's expected to release in the fall. That's been received pretty positively as far as I'm aware, so I don't think that the inclusion of such an “AI” would be a problem to anyone.

I do agree with you here, and I realize that I had made some mistakes in my original post ;-]

However, I still feel the same about generative AI and its big risks to the site. As a personal anecdote, I remember in 2020 getting my siblings into Scratch and watching them make games without any standards, and I remember doing the same as well. When you're a little kid, you don't need a generative AI to make “good art” for you. Having that passion and determination to create is a mindset that shouldn't be interrupted with an AI pulling all the work. I think the only reason why I ended up drawing digital is because of Scratch and that un-interruption.

The face sensing plugin is an amazing, non-intrusive way to introduce AI into coding though. I haven't thought of anything negative to say about it since it came out back in 2022(?).

Here are some block suggestions:

glide (10)

(cloud list ::cloud list)

<sprite[ me v] clicked?>

add (1) to counter

remove (1) from counter

reset counter

(counter)

Write (hi)

<[ v] recieved?>

go to previous position

create clone at x:() y:()

That's it

Reply to the post above :
This topic is not for suggestions, to suggest, make a new topic in the suggestions forum : https://scratch.mit.edu/discuss/1/

Donut1673 wrote:

Here are some block suggestions:

glide (10)

(cloud list ::cloud list)

<sprite[ me v] clicked?>

add (1) to counter

remove (1) from counter

reset counter

(counter)

Write (hi)

<[ v] recieved?>

go to previous position

create clone at x:() y:()

That's it

Hope they add more math functions, like exponents and built-in formulas (such as pi, golden ratio, etc.)

Donut1673 wrote:

Here are some block suggestions:

glide (10)

(cloud list ::cloud list)

<sprite[ me v] clicked?>

add (1) to counter

remove (1) from counter

reset counter

(counter)

Write (hi)

<[ v] recieved?>

go to previous position

create clone at x:() y:()

That's it

1. how would the ‘glide (10)’ block work?
2. cloud lists are rejected (see this post)
3. there's already a workaround for it:

when this sprite clicked
set [clicked? v] to [1] // sprite-only variable?
broadcast [something v]
set [clicked? v] to [0]

([clicked? v] of [somesprite v]) // use this in another sprite

4-7. variables do this job way more simply; also would the counter be global or sprite-only?
8. could use the say/think blocks for this; could also use the animated text blocks from scratch lab; would be a cool addition
9. also rejected
10. also has a workaround:

set [oldx v] to (x position) // sprite-only variable
set [oldy v] to (y position) // sprite-only variable
// ^ set variables before moving
go to x: (oldx) y: (oldy)

11. also has a workaround

set [cid v] to [something] // sprite-only variable
create clone of [myself v]

when I start as a clone
if <(cid) = [something]> then
go to x: (...) y: (...)
end

this block needs to exist

<message(message name v) broadcasted? :: events>

DaSupaOrenFan313 wrote:

this block needs to exist

<message(message name v) broadcasted? :: events>

Rejected

The Official list of rejected suggestions wrote:

1.1 “Broadcast received” boolean block
This in theory would allow a project to detect when a broadcast is sent, but there is ambiguity on how this would work. Would it return true if the broadcast was run at any point after the green flag was clicked, or only on the instant the broadcast was run, or something else? The workaround is simple: use variables that change when a broadcast is received, then use the “equals” block.