How Should Companies Choose Programming Languages?

TL;DR

Companies should choose programming languages by prioritizing product delivery, business survival, and sufficient technical fit over identity or ego. A good-enough language can help a company reach the stage where technical decisions truly matter, while an unnecessary rewrite can destroy velocity, increase spending, shorten runway, and cause the business to miss its market opportunity.

Transcript

There are so many incredible programming languages to pick from nowadays. From Rust to Go to Elixir to, of course, TypeScript, which is why it's kind of crazy that developers are so bad at picking languages. I feel like every conversation I see about languages is either toxic or wrong or some chaotic combination of the two. Why is it like this? I j... Read More

Key Insights

• Programming-language decisions are often shaped by identity, emotion, and ego rather than purely technical analysis, which can turn ordinary engineering discussions into toxic or misleading arguments that overlook the company’s product, financial, and market constraints.
• A programming language is an important company decision, but it is not necessarily the single most expensive choice a company makes. Language choices can also be changed later when the expected operational or development benefits justify the transition.
• A good-enough technology choice can be correct for an early company if it helps the product reach meaningful scale. Twitch’s use of Ruby on Rails had limitations, but it supported the company until moving the front end to TypeScript with React and the back end to Go became worthwhile.
• The key to early company decision-making is reaching the point where technical choices materially affect success. Optimizing prematurely for the theoretically correct language can consume time and money before the business has proven that its product or market opportunity deserves such investment.
• An unjustified rewrite can be more damaging than the original language choice. Tackle’s move from PHP to Perl required both learning a new language and rebuilding the system, causing velocity to collapse and delaying the product by nine months.
• Technical excellence cannot recover lost market timing by itself. Tackle produced a system the author considered beautiful, but it arrived after Facebook expanded beyond colleges, when the company’s opportunity had diminished and its remaining runway was insufficient.
• Runway is a core engineering constraint because higher development costs reduce the time available to find product traction. Tackle’s monthly burn increased from $200,000 to $500,000 after it expanded the team to compensate for reduced velocity.
• Fundraising depends on evidence that a company could become highly successful, such as strong user growth or another promising upward trend. Rapidly increasing engineering expenses caused by a technology rewrite do not provide the momentum required to attract further investment.

Questions & Answers

Q: How should a company choose a programming language?

A company should evaluate whether a language is technically sufficient, supports timely product delivery, and fits its financial runway and current business stage. The goal is not necessarily to select the theoretically best language from the beginning. A good-enough choice may be preferable if it helps the company gain users and reach the point where improved technical architecture produces meaningful business value.

Q: Why do programming-language debates become irrational?

Programming-language debates become irrational when developers treat language preferences as part of their identity and allow emotion or ego to replace a grounded assessment of business needs. The result can be toxic discussion or technical claims that ignore delivery speed, existing team knowledge, rewriting costs, market timing, and the company’s ability to keep funding development.

Q: When is changing programming languages worthwhile?

Changing languages can be worthwhile when the current stack creates meaningful limitations and the expected improvement justifies the transition cost. Twitch moved from Ruby on Rails to TypeScript with React for the front end and Go for the back end. The transition was rough and meaningfully different from starting with Go, but it produced significant improvements and was ultimately considered worthwhile.

Q: Why can a good-enough programming language be the right choice?

A good-enough programming language can be the right choice because an early company first needs to build, launch, attract users, and survive. Twitch’s original Ruby on Rails choice had several shortcomings, but it helped the company reach a stage where those shortcomings mattered. Selecting a more technically appropriate language at the start would not necessarily have produced a better business outcome.

Q: What happened when Tackle switched from PHP to Perl?

Tackle’s new CTO declared PHP to be the wrong choice and ordered a switch to Perl after what the author viewed as an insincere comparison. The engineering team then had to learn Perl and rebuild the system from scratch. Velocity collapsed, the launch was delayed by nine months, monthly spending increased from $200,000 to $500,000, and the company’s runway was halved.

Q: Why did Tackle fail despite building a strong new system?

Tackle completed a system that the author considered beautiful, but the company delivered it too late. Facebook had expanded beyond colleges, the available market opportunity had diminished, and Tackle was approaching the end of its financial runway. The new product lacked enough momentum to meet fundraising milestones, so its technical quality could not compensate for lost time and increased spending.

Q: How does a rewrite affect a startup’s financial runway?

A rewrite can shorten runway by reducing delivery velocity while increasing staffing and operating costs. At Tackle, the company more than doubled its team to compensate for the slowdown caused by learning Perl and rebuilding the platform. Monthly burn rose from $200,000 to $500,000, which cut the runway in half and left less time to achieve the growth required for another funding round.

Q: What evidence helps a company raise more investment?

A company needs evidence that it has a credible chance of becoming highly successful over time. The transcript identifies strong user growth or another sharply rising business measure as the kind of momentum that can support fundraising. An upward curve in expenses, especially after hiring more engineers to recover from a poor technical decision, does not demonstrate the same investment potential.

Summary & Key Takeaways

• Programming-language debates often become irrational because developers connect technologies with identity, emotion, and ego. Although language selection can materially affect engineering work, companies should focus first on reaching the point where technical choices matter. A practical language that enables product delivery may be more valuable than a theoretically better option that causes delays.

• Tackle replaced PHP with Perl after a new CTO conducted what the author described as a sham analysis. The team learned a new language and rebuilt its system from scratch, delaying the product by nine months. Monthly spending rose from $200,000 to $500,000, velocity collapsed, and the company’s financial runway was cut in half.

• The rebuilt Perl system was technically impressive, but it launched after Facebook had expanded beyond colleges and the market opportunity had weakened. Without sufficient momentum, Tackle could not reach the milestones needed to raise more money. The example shows why technical quality cannot compensate for missed timing, excessive spending, or inadequate user growth.