Efficient scaleups in 2024 vs 2021: Sourcegraph (with CEO & Co-founder Quinn Slack)

In 2022 I joined my previous employer, a fintech unicorn, I first learned about Sourcegraph - the company had a subscription, and some (but not many) engineers liked it, saying it was better than the search in GitHub. I found this to be true. I was impressed. In 2023, the free money was gone. The company surveyed the usage of Sourcegraph and cancelled the subscription. Me and many of colleagues were laid off too. However, I have been thinking about this product and even tried Cody when it was released.

The problem with Sourcegraph is that it doesn't move the needle. It's nice to have and convenient, but just not essential enough to become something developers can't live without. Once GitHub improved its search function, there was even less reason to pay for it. Then came AI, and it's just a matter of time before the whole code repository can be placed in context, and before AI can index/retrieve fast enough to provide a very different development experience. It is already moving the needle.

In the next 5 years, will developers still need to write unit tests or even generate them? Probably not. AI will generate unit tests in real time during the build process. If runtime data is made available, AI can even run regression testing and debugging for us. Software engineers will never be replaced, but how they operate will be different - they will be more of a mix of design and review than development.

Therefore, the tooling stack needs to change as well. For example, our current IDEs have been largely the same for 20 years; they are not designed to make reviewing easier. The tooling will need to make the feedback loop much faster than today.

Sourcegraph has a slight advantage right now with their superb knowledge of searching, but they need to think differently about the future development environment and process.

NGL, this was tough to upvote. I did upvote for the effort, but I'd say that going forward the scope of an interview should be more narrow, go deep on a smaller subset of topics.
IMO topics like "hiring new grads" and broad strokes on AI use in the industry should either be avoided or explored separately in a different video or even a different format.

would love to hear more about exactly what scaleups they did and what they regret. Like would love if they can answer more specific questions about their tech stack and stuff. Good video tho!

I would love to work at a company where the CEO Codes.