@curved-petabyte-84246 this is not a very often comparison. But I can share my thoughts.
Temporal is a pretty cool product and I really like it. But Flyte and temporal are dramatically different. Yes they have workflows, but thats where the comparison ends.
Temporal Primary goal - Microservice orchestration, with DIY infra management for the users
Temporal is meant for micro-service orchestration. Things where you want fast moving workflows, but manage the infrastructure on your own, usually services. Focus is not on shipping software in a jiffy, but more on building software products.
Flyte Primary goal: ML Infrastructure with rapid interation and managed infrastructure focus.
On the other hand, Flyte is designed for (currently) slower longer running data and ML workflows, that need recoverability, versioning, reproducibility and most importantly the complete infrastructure is managed.
Think about MLE's, Data Engineers and Data scientists or Software engineers who are doing AI/ML. They do not really want to manage infrastructure. They do want reliable executions, but often times shipping faster is more important. Shipping from local to remote is optimized (and more will be).
Flyte is also inherently infrastructure aware - know gpus, cpus etc (soon also will know fractional gpus), also knows quotas and can apply backpressure
Flyte eventough language agnostic, not really the same as temporal