Amjad Masad, CEO of Replit: How Bounties Will Change the Future of Software Engineering
Bounties, AI, remote work, and new tools will transform how software is built and change the part-time work landscape
Amjad Masad is the CEO of Replit, “The Collaborative Browser-Based IDE”. Founded in 2016 and backed by Y Combinator, they’ve raised over $100m on their way to building a great product. I’ve used it myself to tinker on projects, launch things quickly, as well as host the website for my farm.
According to Amjad though, the IDE branding is outdated as their visions have become much grander. How does this relate to Part-Time Tech?
Replit recently announced their Bounty Program where anyone can post a software bounty to be completed by someone on the platform. While still a very early product, I think bounties represent one glimpse of what the future of software engineering and part-time work could look like.
Naturally, when most people think of working part-time, they think of their full-time job done in <40 hours/week. While I think this type of role will become more prevalent over time, it’s far more likely that most part-time work will be a smorgasbord of new types of work that never existed before, such as bounty work for a company across the world because the workflows are embedded right into their IDE.
Often times when we look at major waves of cultural or technological change we expect it to fit the mold of our past frame of reference (e.g. part-time work will be full-time work with fewer hours) when chances are a much larger shift and repositioning is yet to occur.
Ok, enough editorializing, let’s jump to the Q&A with Amjad.
Replit is branded as the collaborative IDE. What’s your long term vision for Replit?
That's somewhat of an outdated branding. Ultimately Replit is about reducing the distance between an idea to a product. In fact, from an idea to your first dollar earned -- how fast can you start making money from an idea. Previously, to write your first line of code you had to wade through a mountain of crap that is setting up the development environment, and then figuring out how to deploy your code and share it is another ordeal. Replit simplifies this down to the essentials. We give you an environment to learn and code, complete with an AI assistant to make you go faster and help unstuck you. We also make it super easy to share code with people, whether collaborating or using your app.
So now you have tools that are a joy to use and it's a breeze to share your code. Now, you need to figure out how to monetize it; yet another adventure. So we're solving that too. We recently introduced Cycles, our in-platform currency. You can earn Cycles in the old-fashioned way by simply doing work, or you can get tips for your apps that people find useful, and finally, pretty soon, you'll be able to charge your users' in Cycles directly.
Replit Bounties is a new project you launched where users can put up bounties for software projects. How do they work and what’s the future vision?
Replit is about making software. You can either learn to code and use our tools to make it, use our AI to make it for you, or hire someone from our community to make it for you. The last case is Bounties. It's very simple: post a description of the thing you want to be built, Replers [Replit users] apply to build it for you, and you pick someone and work with them to get it done. Most bounties are done in a matter of hours to days.
Folks getting the most out of bounties today are our tech professionals that don't know how to code or can't be bothered to do it. We essentially found a natural trade on our platform. We have people with a lot of money who want work done, and then we have people with a lot of time who want work to do. We're connecting them.
Since 2016, Replit has helped millions to learn to code.
Today, we're showcasing how people are using that code to earn.
Welcome to the newest digital economy.
— Replit ⠕ (@Replit)
Dec 15, 2022
Bounties, while early now, seems like a very different way to work in a globalized, remote, tech economy. How do you envision Replit changing how people work in the future?
Long term, we want to reduce the transaction cost of software down to near zero.
In other words, outsourcing software or finding an expert to solve a specific problem should be as easy as getting an Uber. Tactically, this means building Bounties right into the IDE and the software-making process. Imagine highlighting a piece of code, right-clicking, and "create bounty," and then you ask for the code to be optimized. A few minutes later, a coder expert at optimization appears in your IDE, writes the code, review it/tests it, and then pays them. And I think this will extend to thinking outside code, like labeling data for ML or doing QA.
Imagine farming out work right from your editor. The consequences of near-zero transaction cost and friction will be insane.
— Amjad Masad ⠕ (@amasad)
Nov 15, 2022
So in the limit, this makes needing a large team to build a super successful company unnecessary and perhaps discouraged. Today, a dozen-person $1B team like Instagram is seen as a rare thing, but in the future, it'll be more common, and we'll see lots of single-person unicorns.
What do you think the software industry looks like in 10 years?
This is best answered in this Tweet thread. (Copied below for readability)
I don't think people understand the monumental changes coming to software this decade. Quick thread:
— Amjad Masad ⠕ (@amasad)
Nov 23, 2022
Last century making software got progressively easier going from machine code to assembly to higher-level and then scripting languages. The last major productivity boost in software was OSS. Each of those steps was 10-100x boost but then it stopped...
AI is the next 100x productivity boost. Copilot/Ghostwriter is just the early innings bringing 30-50% improvement. The next generation coding AI will not be mere text complete and will lead to rapid change in how we make software.
At Replit, we're building an AI pair programmer that uses the IDE like a human does and has full access to all the tooling, open-source software, and the internet. In the next few years, programmers will operate at a higher level than mere code.
Crucially, it won't be "prompting" -- we believe that's more a bug than a feature -- it will be a combination of the AI predicting what task you want done next and doing it for you, plus a dialog-based agent that follows your commands.
Programmers will command armies of software agents to build increasingly complex software in insane record times. Non-programmers will also be able to use these agents to get software tasks done. Everyone in the world will be at least John Carmack-level software capable.
The other seismic shift will be coordination primitives for developers. Chief among them is payment primitives. Bitcoin Lightning, for example, bakes value right into the software supply chain and makes it easier to transact both human-to-human and machine-to-machine.
Driving the transaction cost and overhead in software down means that it will be a lot easier to bring developers into your codebase for one-off tasks. Lightning helps with coordination, where staking can keep participants honest and pay for the exact work.
One way to visualize this is that software will move from a stack to a network model. In the stack world, we assemble code in a repo and ship it somewhere to run and then monetization is bolted on. In a network model, code is fully monetized and running all the time.
Put these things together, one developer will have the power of an entire network of AIs, people, and services at their fingertips. I believe a 100x productivity boost is the lower bound here.
How can bounties expand to encompass larger work? Right now the projects are fairly small in scope and low cost.
Individual bounties don't need to be large work for them to power large projects. I think the future is a mixture of AIs and people working together to build large projects. The collaboration could be ad-hoc or asynchronous or teams could be formed on the fly by bounty hunters pooling together to form teams that would disband after the project. This could also all be automatically orchestrated.
Could there be a race to the bottom where bounties get progressively cheaper as talent in lower cost labor areas scoop up these projects?
Like in any market, that will happen for low-skill work. But high-skill work -- e.g., performance optimization -- it will likely get more expensive/lucrative.
Are you going to move into credentialing and proof of work based on bounties and other Replit portfolios?
It should happen informally initially, and then we can productize. We'll see bounties on resumes as early as in the next few months.
Do you expect large, established software organizations incorporating bounties into their workflows?
Yes, as Replit gets adopted in the enterprise. They'd want better security & legal guarantee, but this will happen.
Do you expect people to be able to make a living off of Bounties or have their entire career be on Replit?
That's already happening for hunters living in low-income countries.
How was this post?
Feedback helps us improve future content.