If that were the case, would developer-lead startups not have a 100% success rate? After all, a function that takes hours to complete, when an optimized function could take milliseconds, is still more than fast enough if you have no users. I'm not sure anyone has actually proven that they can make such predictions accurately.
No, Pro Poker player loose all the time, but they are far better than your average players becasu they make better bets, because they understand the game and the odds better. Good programers are wrong all the time, (I know i am) but they make less misstakes, and they can fix the misstakes faster, because they can predict when there may be an issue.
Also being a good programmer is not the same as being good at running a startup.
> Also being a good programmer is not the same as being good at running a startup.
But being able to predict the future of the business is necessary to determine where to optimize. Again, if you have no users, you don't need to optimize at all –period. The success rate should be 100%, as those who predict what to optimize will know when not to go into business.
If you get it wrong, you haven't predicted anything. You were guessing.
1. GP didn't say that they could predict what will be "fast enough" they can predict what will be "slow".
2. A kind reading of GP would also understand that one is predicting what will be slow under some set of conditions.
3. Requiring 100% accuracy to call it a prediction is absurd. A weather forecast is definitely a prediction even though it's not 100% accurate. The NOAA isn't "guessing" weather in the general usage of the term "guess"
4. By your definition extensive benchmarking, measuring a program, and optimizing based on those results is wrong, because we don't even know if the end user will ever run the program!
If that were the case, would developer-lead startups not have a 100% success rate? After all, a function that takes hours to complete, when an optimized function could take milliseconds, is still more than fast enough if you have no users. I'm not sure anyone has actually proven that they can make such predictions accurately.