This looks neat to me, but I'm not really sure what advantage this distribution model offers over their current offerings. For one, every user is going to need some way to execute your arbitrary code, meaning most will need a sandbox at the very least, all the way up to a small dedicated VPS for beefier programs. Of course this begs the question, will there be a pricing model? Will certain languages be more expensive to run than others? Will some languages be outright omitted? Will each execution of the program also include compilation time?
Hats off to the repl team for making this work, but there's a lot of roadblocks ahead. I'm sure they have the immediate capital to make it work in the short-term, but I struggle to see how this would be a sustainable project...
Does replit have a really high cadence compared to most startups or am I surprised because I just see the occasional announcement and these things brew for a long time under the radar?
My opinion is repl.it iterates fast, the CEO founder is very shipping and community oriented, and he's a good follow on Twitter: https://twitter.com/amasad
They were/are also incubated by YCombinator. They’re a darling here, for all the reasons you mentioned, but I also get the impression that they build in a “see what sticks” manner. I’ve found some of their recent announcements neat but not immediately valuable long-term or for non-casual users.
I have also been very impressed by their speed and would like to know their secret. They were founded in 2016 and I have only noticed this speed of iteration in the past 6-12 months. Was there an actual uptick? And was there some work or infrastructure investment in their early years that is now paying dividends?
I agree. It seemed like another competitor in the space along with cloud9, koding, code anywhere, shiftedit, code envy, nitrous and who knows how many others founded earlier, and then it seemed to move away adding on features.
Really the big moves are towards hosting, and I always thought that after AWS bought cloud9, they would dominate that space, but cloud9 seems frozen.
I wonder if Digital Ocean or someone buys them to bring about the cloud IDE/hosting combo. (Actually, I have no idea how large replit is and if that is even reasonable).
Maybe I'm just naive, but I find this insanely exciting. Particularly the analogy of Youtube vs TikTok. Having the source code right next to the executable, right there in your browser seems very powerful.
If it pans out right, the next gen of memes could be games like the ones showcased here, and then people "fork" it, and change some of the code and make it their own.
I have nothing to do with repl.it, just really happy to see this.
I remember doing this in high school manually. We took a fly swatting game and replaced the sprites with my friends face (regular and splatted). A platform that enabled all of this would be out of control
This is really interesting because this is what glitch.com seemed to promote as their number one feature for awhile. From their latest homepage it seems like they took a step back and are promoting a web-based IDE with the forks being a little hard to find.
We’re always iterating, testing out different emphasis on creation vs. discovery since we care a lot about both. But every one of the 10 million+ apps on Glitch already has a page where it lives so folks can remix it or share it with their team or add it to a collection.
So if you take something like the official Eleventy starter, it has a page like this, and you can see who else has added it to a collection, etc. https://glitch.com/~11ty
Generally people are finding Glitch apps through their peers who make them, or embedded in the docs they read, or by someone sharing it at a hackathon or whatever, so we de-emphasized direct discovery on the homepage. If folks like being able to find things that way, I’m sure we’ll bring it back; it’s not a comment on the utility of having the ability to create, share & reuse millions of web apps easily, obviously we’re hugely in favor of that.
Is it another attempt to "make programming easy"?. I keep wondering if it's even possible. Can everyone be code literate? or do you have to have an "algorithmic" (for lack of a better word) way of thinking? Is it something like woodworking class vs real carpentry?
Minecraft, for example, opened a door to system thinking to a lot of young people, but there's probably many more that just like the fun and not the understanding and craft, so maybe those people's brain already leaned towards it?
Right now we're in peak "Everyone should know how to code" but sometimes it seems like we keep chasing the golden goose - Visual Basic, Dreamweaver, Node based programming, Webflow etc.
Edit: forgot to add that the CEO has that confident delusional/visionary vibe so this certainly fits the bill. Time will tell which it is.
I agree that not everybody needs to code. I disagree with the implication that the current ecosystem isn’t excluding some people who could code: it is.
Frankly, a lot of people who currently write code are only writing code because they stumbled across the right combination of tools: reducing the barrier to entry will only improve things.
A lot of people know basic HTML/CSS and the next generation is learning it in their computer classes, but there aren’t many tools that let them leverage it to build actual websites. If there were, people wouldn’t be so forced into proprietary, expensive solutions. And a lot more people would realize code isn’t that hard. That’s what we’re trying to do with primo (https://primo.af).
Why not, for the same reason that everyone can be - to reasonable standards - be literate?
This is merely a new and powerful form of literacy that has become relevant due to a particular set of technological circumstances, but structurally it's not that different from the printing press preceding widespread literacy of the regular sort.
And no doubt this form of literacy is powerful. It relates to an understanding of how to describe the processing of information in various contexts. In an increasingly information overloaded world, the ability to use the tool of your time to process the information you need will become a critical ability.
I think one fallacy we can implicitly succumb to is the notion that "everybody knowing how to program" somehow relates to everyone being a programmer.
Everyone knowing how to read and write doesn't mean everyone is an author or a technical writer. But everyone _does_ use their ability to read and write to make their own work more productive within their contexts. And even in a personal context, being able to read an ingredients list, or write a shopping list, is underpinned by literacy.
What's so hard to imagine about a future where everyone knows how to build small useful programs for themselves, and just as everyone has access to post it notes and notepads, everyone having access to a few small devices that allow them to use that literacy to build programs that are useful to them in their day to day activities?
It's hard to imagine the specifics of how that might be applied, but it seems easy to imagine the transformative power of that sort of a casual power being not just available to the masses, but understood and able to be exercised by all of them.
> Why not, for the same reason that everyone can be - to reasonable standards - be literate?
Not everyone can become literate. Very roughly 0.1% of people lack the mental ability to read, even given lots of training. Flipping your argument around, it's totally possible that programming is that much more difficult than reading that 0.1% of the population isn't unable, but more like 10%.
It’s not an agreement that it’s practically achievable any time soon. It’s just that more than 90% of people have brains that are capable of literacy (or did at some point).
I think the problem is equating "coding" to "algorithmic thinking."
Most coding problems are wiring stuff up. If you can change a tire or put together a LEGO set, you can probably write code.
The problem with the golden geese that you're pointing out is mostly one of tooling, in my opinion. Undergrads at top CS programs are as clueless as your grandmother about developing software because the tools are either toys for children or extremely sharp knives with a blade on the handle. This is a problem that replit is trying to solve.
I wish spreadsheets had better support for version control and automated regression testing. Tons of people can program, and they'd all be well-served by such tools.
I think everyone should be able to write a Python script to change a CSV file. Likewise with reading, everyone can read stuff on Instagram. Most of us can't read the Magna Carta.
Likewise I'm pretty good but I can't code in Assembly
I started coding as a kid with VB 6, then switched to VB.Net. I don't remember that phase all that well (it was years ago), but I think more mainstream languages are actually easier. One thing that was great about VB was Visual Studio's great GUI, GUI creator. I have still not seen anything as simple, convenient, and productive for (simple) UI development as Windows Forms in Visual Studio.
Lundum Dare 48 was a couple weeks ago and since I was aware of the jam from its promotion on pygame.org and I'm a python guy I wanted to make something in pygame. Come to find out, replit has a pygame enabled environment.
It's not very fast. It worked for some simple animations and a stupid little turn based colony building simulator but I dunno if it could handle something more arcade style.
I wouldn't call the IDE great. It's all on the web and subject to unsurprising latency. I ran into a couple of weird bugs.
But I spent zero time setting up my environment. Zero time building binaries. Zero time on cross platform testing. Publishing (to the jam) was literally pasting the link to the replit on my jam page. Everybody got my source code and a place to hack on it. That is pretty damn cool.
Awesome Brian! I love hearing stories like this. And interesting game. I love the stick-figure animation.
>I wouldn't call the IDE great. It's all on the web and subject to unsurprising latency. I ran into a couple of weird bugs.
We have a big IDE quality sprint coming -- we just hired amazing engineers fully focused on it (and still hiring https://replit.com/careers) and we're going to be making big gains soon :D
You can always email me feedback directly as well amjad@replit.com
how does Replit Apps compare to Glitch? i feel like there are so many similar attempts to crack this nut and while the choice is nice, i remember being a little overwhelmed when i was starting out.
47 comments
[ 4.0 ms ] story [ 101 ms ] threadWith an app store and server infrastructure with public ports, I could also see API/automation browsing built-in.
And hasn't nearly every strategy in the history of computing that started with gaming succeeded wildly?
Hats off to the repl team for making this work, but there's a lot of roadblocks ahead. I'm sure they have the immediate capital to make it work in the short-term, but I struggle to see how this would be a sustainable project...
Pricing is here: https://replit.com/site/pricing
Really the big moves are towards hosting, and I always thought that after AWS bought cloud9, they would dominate that space, but cloud9 seems frozen.
I wonder if Digital Ocean or someone buys them to bring about the cloud IDE/hosting combo. (Actually, I have no idea how large replit is and if that is even reasonable).
If it pans out right, the next gen of memes could be games like the ones showcased here, and then people "fork" it, and change some of the code and make it their own.
I have nothing to do with repl.it, just really happy to see this.
Even further: Dub Music https://en.wikipedia.org/wiki/Dub_music#%22Versions%22_and_e...
So if you take something like the official Eleventy starter, it has a page like this, and you can see who else has added it to a collection, etc. https://glitch.com/~11ty
Generally people are finding Glitch apps through their peers who make them, or embedded in the docs they read, or by someone sharing it at a hackathon or whatever, so we de-emphasized direct discovery on the homepage. If folks like being able to find things that way, I’m sure we’ll bring it back; it’s not a comment on the utility of having the ability to create, share & reuse millions of web apps easily, obviously we’re hugely in favor of that.
Minecraft, for example, opened a door to system thinking to a lot of young people, but there's probably many more that just like the fun and not the understanding and craft, so maybe those people's brain already leaned towards it?
Right now we're in peak "Everyone should know how to code" but sometimes it seems like we keep chasing the golden goose - Visual Basic, Dreamweaver, Node based programming, Webflow etc.
Edit: forgot to add that the CEO has that confident delusional/visionary vibe so this certainly fits the bill. Time will tell which it is.
Frankly, a lot of people who currently write code are only writing code because they stumbled across the right combination of tools: reducing the barrier to entry will only improve things.
I never used any tools like Replit when I was learning to code.
Why not, for the same reason that everyone can be - to reasonable standards - be literate?
This is merely a new and powerful form of literacy that has become relevant due to a particular set of technological circumstances, but structurally it's not that different from the printing press preceding widespread literacy of the regular sort.
And no doubt this form of literacy is powerful. It relates to an understanding of how to describe the processing of information in various contexts. In an increasingly information overloaded world, the ability to use the tool of your time to process the information you need will become a critical ability.
I think one fallacy we can implicitly succumb to is the notion that "everybody knowing how to program" somehow relates to everyone being a programmer.
Everyone knowing how to read and write doesn't mean everyone is an author or a technical writer. But everyone _does_ use their ability to read and write to make their own work more productive within their contexts. And even in a personal context, being able to read an ingredients list, or write a shopping list, is underpinned by literacy.
What's so hard to imagine about a future where everyone knows how to build small useful programs for themselves, and just as everyone has access to post it notes and notepads, everyone having access to a few small devices that allow them to use that literacy to build programs that are useful to them in their day to day activities?
It's hard to imagine the specifics of how that might be applied, but it seems easy to imagine the transformative power of that sort of a casual power being not just available to the masses, but understood and able to be exercised by all of them.
Not everyone can become literate. Very roughly 0.1% of people lack the mental ability to read, even given lots of training. Flipping your argument around, it's totally possible that programming is that much more difficult than reading that 0.1% of the population isn't unable, but more like 10%.
Most coding problems are wiring stuff up. If you can change a tire or put together a LEGO set, you can probably write code.
The problem with the golden geese that you're pointing out is mostly one of tooling, in my opinion. Undergrads at top CS programs are as clueless as your grandmother about developing software because the tools are either toys for children or extremely sharp knives with a blade on the handle. This is a problem that replit is trying to solve.
Likewise I'm pretty good but I can't code in Assembly
It's not very fast. It worked for some simple animations and a stupid little turn based colony building simulator but I dunno if it could handle something more arcade style.
I wouldn't call the IDE great. It's all on the web and subject to unsurprising latency. I ran into a couple of weird bugs.
But I spent zero time setting up my environment. Zero time building binaries. Zero time on cross platform testing. Publishing (to the jam) was literally pasting the link to the replit on my jam page. Everybody got my source code and a place to hack on it. That is pretty damn cool.
For the curious: https://replit.com/@BrianDavis7/A-Disguised-Far-Muse
>I wouldn't call the IDE great. It's all on the web and subject to unsurprising latency. I ran into a couple of weird bugs.
We have a big IDE quality sprint coming -- we just hired amazing engineers fully focused on it (and still hiring https://replit.com/careers) and we're going to be making big gains soon :D
You can always email me feedback directly as well amjad@replit.com
- more modalities: you can make CLI apps, web apps, even native apps (like PyGame)
- more language support
- more complete platform: auth, db, etc
- bigger community
On Apps more specifically, the hashtag, trending, and the general exploration UX aspect is going to be huge for communal creativity.
Seriously, I think this is an exciting idea. There are probably a fraction of a billion people who know how to code, even if not experts.
This reminds me of my granddaughter writing games with a game system from MIT (which I forget the name of) and sharing them with friends.