Elon Musk’s Twitter RPC War: Engineers Clash Over 1200 vs 200 Microservice Calls

When Elon Musk posted an apology tweet saying Twitter suffers from ultra‑slow speeds in many countries and that the app is executing over 1,000 improper batch processes with countless RPC calls, a wave of engineers immediately challenged his numbers.

Technical lead Sach (Sachee) replied that Musk does not understand the technology, calling his remarks rude and claiming Musk is ignorant of GraphQL. She posted a screenshot criticizing his lack of technical knowledge.

Engineer Eric Frohnhoefer entered the debate, stating that Twitter actually performs zero RPC calls when the app starts, only about 20 backend requests. He later clarified that the number of RPC calls needed to generate the home‑timeline is closer to 200, not the 1,200 Musk mentioned.

Musk responded by insisting that up to 1,200 microservices can be invoked for a single request, while Frohnhoefer maintained his lower figure. The back‑and‑forth became chaotic, with multiple fragmented statements across timelines.

Both Frohnhoefer and Sach were eventually fired by Musk, as confirmed by their own tweets. External commentator Sam Pullara, a former Twitter and Yahoo engineer, explained that the real slowdown stems from the removal of server‑side rendering, forcing users to download large bundles of code, especially in regions with poor network conditions.

Twitter later announced plans to shut down a portion of its outdated microservices, noting that the platform now needs less than 20% of the time it previously required to operate.

Frohnhoefer posted on LinkedIn that the experience made him “more popular than ever,” while a Reddit developer publicly invited him to apply for an Android senior engineering role, highlighting the community’s support despite the internal turmoil.