In the 2024 annual report (released today), the message from Scratch's President/CEO has this:

As we look ahead to Scratch 4.0 and our ambitious goal to reach the next 100 million children by 2027, …

Also, 4.0 was teased earlier in a Scratch job application section (which can't be linked because it indirectly leads to filling out contact information):

In 2025, we’re responding to this growth by focusing on three strategic priorities: developing the new Scratch 4.0, with a re-imagined online community and creative AI; amplifying the impact of Scratch through our programs and research; and developing a robust business model to ensure a sustainable future for the software and the Foundation.

It was announced back in April that AI would be added to Scratch, so it seems 4.0 may be coming in the not-so-distant future. Any thoughts?

I'm excited : )
also st please update the profile page and forum theme to 3.0 at least

Did they also leak 3.0 like this? Or did it just come out of nowhere?

Edit: waittt…… AI!?!!?! for Scratch!?!! Yes, AI is the hot new tech right now, but really, it has specific applications, and telling kids how to make games and animations and making assets for users is not one of them. Scratch is for creativity and not making ai do the work for you (no creativity there) .

minniesworld wrote:

Did they also leak 3.0 like this? Or did it just come out of nowhere?

This isn't really a leak, per-say. More like a semi-announcement.

As for if they announced it in this way. Maybe? They had a period where 3.0 was in a beta period, but I'm not sure what was up with it before then.

—

I don't feel like they should be focusing on creating a new version of Scratch, but rather in putting more time into 3.0. 3 is already extremely versatile, so they should be improving some of its flaws rather than creating a new version from scratch for a 3rd time. Unless they're planning on just updating 3.0 to make this “4.0.”

I really can't think of anything that would really justify making a full-on 4.0. Development on 2.0 was started because 1.0 was made in a programming language that wasn't that good for the web, and 3.0 was started because 2.0 was made with Flash, which was (even back when 2.0 was new), really old and filled with security holes. But 3.0? 3.0 doesn't (to my knowledge) have any major security holes nor is it written in a language that isn't very good for the web. 3.0 was written entirely in JavaScript, which is perfect for the web; 3.0 also doesn't have any major security holes, as far as I am aware.

Having a newfangled “Scratch 4.0” just seems entirely unnecessary, at least in my opinion.

Only really big thing I can really think of that would even somewhat justify a re-write, would be how incredibly divided the whole codebase is, and wanting to unify them a lot more, like how 2.0 and 1.4 were. The codebase may be an internal development nightmare, given the way that it's arranged.

—

As for my thoughts on AI, I don't really care for it. As long as it's entirely optional, I won't mind. But also I really wouldn't want to see projects made entirely with AI. At a certain point, stuff goes from “being creative” to just having something make something on your behalf, which doesn't really encourage learning. Studying stuff can help with learning a skill, but generating a full game at looking at the spaghetti the AI spits out really doesn't help in that respect, especially when you have no knowledge of what anything actually does.


- Nedroid's “The Internet,” 2013

Steve0Greatness wrote:

minniesworld wrote:

Did they also leak 3.0 like this? Or did it just come out of nowhere?

This isn't really a leak, per-say. More like a semi-announcement.

As for if they announced it in this way. Maybe? They had a period where 3.0 was in a beta period, but I'm not sure what was up with it before then.

—

I don't feel like they should be focusing on creating a new version of Scratch, but rather in putting more time into 3.0. 3 is already extremely versatile, so they should be improving some of its flaws rather than creating a new version from scratch for a 3rd time. Unless they're planning on just updating 3.0 to make this “4.0.”

I really can't think of anything that would really justify making a full-on 4.0. Development on 2.0 was started because 1.0 was made in a programming language that wasn't that good for the web, and 3.0 was started because 2.0 was made with Flash, which was (even back when 2.0 was new), really old and filled with security holes. But 3.0? 3.0 doesn't (to my knowledge) have any major security holes nor is it written in a language that isn't very good for the web. 3.0 was written entirely in JavaScript, which is perfect for the web; 3.0 also doesn't have any major security holes, as far as I am aware.

Having a newfangled “Scratch 4.0” just seems entirely unnecessary, at least in my opinion.

Only really big thing I can really think of that would even somewhat justify a re-write, would be how incredibly divided the whole codebase is, and wanting to unify them a lot more, like how 2.0 and 1.4 were. The codebase may be an internal development nightmare, given the way that it's arranged.

—

As for my thoughts on AI, I don't really care for it. As long as it's entirely optional, I won't mind. But also I really wouldn't want to see projects made entirely with AI. At a certain point, stuff goes from “being creative” to just having something make something on your behalf, which doesn't really encourage learning. Studying stuff can help with learning a skill, but generating a full game at looking at the spaghetti the AI spits out really doesn't help in that respect, especially when you have no knowledge of what anything actually does.


- Nedroid's “The Internet,” 2013

I highly doubt that just rewriting Scratch's code, even for the full program, would warrant a “4.0” label, as this would only disappoint children who were expecting something major to change. I agree that Scratch 3.0 doesn't really need an upgrade as big as “4.0”, and the Scratch Team seems to agree. The code is a little spaghetti and there are a few crashes and security holes from time to time, but it works “well enough” and the flaws are not very noticable for the average user.

I'm personally still a bit skeptical that a “4.0” will happen anytime this decade. Scratch 3.0 is being updated more frequently compared to the pre-purple era (accessibility update, 3.0 scratchblocks forums update, new starter projects page, etc.), and it seems like the Scratch Team just wants to make updates in batches and implement them into Scratch 3.0 whenever they're ready (especially with the way Scratch Labs is set up, which is where their AI experiments currently live). I definitely would like to see a 4.0 update happen as there's a lot of improvement that can be made to Scratch, but I'm not sure that that's going to happen anytime soon.

I've been seeing “NGP” (Maybe it stands for Next Generation Platform?) on some pull requests for the scratch-editor repo. They seem official. Could Scratch 4.0 be referring to a redo of Scratch as a platform?

Edit: It seems like a lot of recent work on Scratch are actually from Astea Solutions, including pull requests related to NGP.

Astea Solutions wrote:

The Scratch Foundation promotes creative coding education by providing free access to the Scratch and ScratchJr platforms, supporting a global community where children can learn to code and express their ideas.

Scratch pioneered block-based coding. Created in 2007 by Professor Mitchel Resnick of MIT and the Lifelong Kindergarten group at MIT’s Media Lab, Scratch is based on decades of educational research. With Scratch, children create their own interactive stories, games, and animations, and share them with friends, teachers, and the online community. Scratch is the world’s largest free coding platform for kids, reaching 20 million new online users each year.

Astea Solutions collaborates with Scratch on design, architecture and development of a next generation Scratch platform, including new backend systems, improvements to the editor and automated inappropriate image detection, while simultaneously improving the current Scratch experience across the board.

The reason I'm quoting them and not linking directly is because the page I quoted it from has contact information.

In my opinion, 4.0 will just be 3.0 if it hasn’t been rushed due to Flash die

Steve0Greatness wrote:

I really can't think of anything that would really justify making a full-on 4.0. Development on 2.0 was started because 1.0 was made in a programming language that wasn't that good for the web, and 3.0 was started because 2.0 was made with Flash, which was (even back when 2.0 was new), really old and filled with security holes. But 3.0? 3.0 doesn't (to my knowledge) have any major security holes nor is it written in a language that isn't very good for the web. 3.0 was written entirely in JavaScript, which is perfect for the web; 3.0 also doesn't have any major security holes, as far as I am aware.

Having a newfangled “Scratch 4.0” just seems entirely unnecessary, at least in my opinion.

Scratch 4.0 will almost certainly be an upgrade to 3.0 and not a complete rewrite. That's how major software updates are typically made – the fact that Scratch has been completely remade twice so far is very unusual.

Based on what I've seen on GitHub, it seems like the're rewriting the website but not the editor. The new website will unfortunately be closed-source.

o97doge wrote:

I've been seeing “NGP” (Maybe it stands for Next Generation Platform?) on some pull requests for the scratch-editor repo. They seem official. Could Scratch 4.0 be referring to a redo of Scratch as a platform?

This PR contains a lot of interesting discussion related to NGP: https://github.com/scratchfoundation/scratch-editor/pull/17

Huh, so it could be possible NGP (which is possibly Scratch 4.0) will support signing in with Google. I see no other reason they would need Google avatars.

o97doge wrote:

Huh, so it could be possible NGP (which is possibly Scratch 4.0) will support signing in with Google. I see no other reason they would need Google avatars.

They might as well implement it- plenty of other websites support it, the community has been asking for it (somewhat, but I'm pretty sure there's a thread in Suggestions for it) for ease of use, and honestly these days it's just something most organizations (including schools) expect you to have.

Although, they better not make those the default profile pictures if you sign up with Google, since many people use images of their face in their Google profile picture, which isn't allowed on Scratch (but new Scratchers wouldn't necessarily know this).

I guess this is what @Jeffalo was talking about when they mentioned that the ST might be working on something to replace scratch-www entirely a few months ago

I don't know how to feel about this, it all seems so gimmicky and corporate.

I agree with @Steve0Greatness , ST should rather put some time into 3.0 instead of yet creating ‘4.0’, also, can’t we just add whatever ST wishes to cook up onto the pre-existing 3.0? It sure would save time and money just focusing on 3.0 for now until there is definitely a sign for 4.0.

23mConnor wrote:

I agree with @Steve0Greatness , ST should rather put some time into 3.0 instead of yet creating ‘4.0’, also, can’t we just add whatever ST wishes to cook up onto the pre-existing 3.0? It sure would save time and money just focusing on 3.0 for now until there is definitely a sign for 4.0.

Though it might be cool to see some sort of Ai though, in my opinion, so long as it is not excessive.

AI is helpful in many ways, but I don't think it should be used too much on a learning oriented platform for kids. In my opinion, the most important thing kids need to learn is creativity. Having an AI draw images, debug code, etc. removes that creative, problem solving aspect from Scratch. Of course, I don't know what AI will be doing, but I can imagine it will do the afore mentioned tasks.

Although the idea of a new Scratch version sounds exciting, I also agree with Steve0Greatness and 23mConnor. Just think of how many times the Scratch servers went down, how many bugs there are on the site, how vulnerable Scratch is to attacks from skibidi farm bla bla bla trolls. Scratch should definitely focus on perfecting what has problems, not what is already good enough.

Why do people put AI into everything? AI is magic and not programming, sometimes even the creators don't know what they put into it / train it on. This isn't safe

nembence wrote:

Why do people put AI into everything? AI is magic and not programming, sometimes even the creators don't know what they put into it / train it on. This isn't safe

AI generated images could be a good way to encourage creativity… if it had super tight safeguards. And I doubt any existing AI is up to the task.

rdococ wrote:

nembence wrote:

Why do people put AI into everything? AI is magic and not programming, sometimes even the creators don't know what they put into it / train it on. This isn't safe

AI generated images could be a good way to encourage creativity… if it had super tight safeguards. And I doubt any existing AI is up to the task.

There is no creativity if you are not creating anything.

NONONONONNONOONONONONONONON WHAAAAAAAAAAATTT GENERATIVE AI TO SCRATCH!?!??!?! I KNOW I MADE A SUGGESTION FOR THIS MYSELF BUT THE WAY THEY SAID IT SEEMS LIKE IT'S GOING TO BE WAY MORE DEEPLY INTEGRATED THAN JUST A HELPER TOOL NOOOOOOOOO!!!! THIS IS LITERALLY THE CREATICODE SITUATION IN SCRATCH!!!! AHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH

Also, why are they saying all this 4.0 stuff now, and working on 4.0? Half the site isn't even 3.0 yet, and it is still 2.0. So they are now working on 4.0? Yes, I looked into the Scratch codebase, mostly for the GUI stuff and see that there's some average coder stuff like arbitrary but necessary values, some inefficiency's, and like 10 million files, some are a whole file or even two for a single button (That's React for you, folks). Given that the Scratch codebase seems pretty spaghetti and messy, it seems reasonable that they haven't even upgraded the scratchr2 parts yet. But I feel like that's what they should be focusing on right now, and just setting 4.0 aside for a moment. Like, the Scratch forums still run on the somehow-still-functioning-despite-ancient DjangoBB, which has been discontinued 10 times over and the last update to the main site was in like, 2009. It has been heavily modified to fit in Scratch, and has had vulnerability fixes to keep it running, leaving Scratch basically one of the only (if there even are any other) functioning and living instances of DjangoBB now. I also heard from somewhere that scratch-www, which is the 3.0 frontend, doesn't even have direct access to the database, and uses the API, which we all know and love, while scratchr2 does have access and is why things on scratchr2 load way faster, and have less terrific loading glitches compared to scratch-www (recently, scratch-www has broken like 3 times for short periods of time). So, yeah, if the Scratch Team ever reads this, just please focus on finishing 3.0 before going all in on 4.0. Just a thought.

N8_D_GR8_1 wrote:

rdococ wrote:

nembence wrote:

Why do people put AI into everything? AI is magic and not programming, sometimes even the creators don't know what they put into it / train it on. This isn't safe

AI generated images could be a good way to encourage creativity… if it had super tight safeguards. And I doubt any existing AI is up to the task.

There is no creativity if you are not creating anything.

you shouldn't have to be a good artist to be able to make the game you want

ispretty wrote:

N8_D_GR8_1 wrote:

rdococ wrote:

nembence wrote:

Why do people put AI into everything? AI is magic and not programming, sometimes even the creators don't know what they put into it / train it on. This isn't safe

AI generated images could be a good way to encourage creativity… if it had super tight safeguards. And I doubt any existing AI is up to the task.

There is no creativity if you are not creating anything.

you shouldn't have to be a good artist to be able to make the game you want

You already don't have to be.