The way I see it, as soon as you think about it seriously it becomes obvious that all (or at least most) the negative feelings we have are there for a very good practical reason. Boredom is effectively a push to prevent an agent from being stuck in a local optimum, and required to get a more effective explore-vs-exploit tradeoff (for whom we have theoretical grounding for what's good or not based on the analysis of "multi-armed-bandit" problems), and the same applies for satiating pleasure e.g. hedonic treadmill; burnout is required for an agent to be able to "say" "this isn't working, you need to do literally anything else"; fear has obvious benefits for risk management; and anger, hate and vengeance has obvious causes in game theory about optimal strategies in games with potential defectors.
So that is an argument that any effective general intelligence is likely to possess at least some of the emotions that we recognize, simply because they are effective for a purpose.
On the other hand, these factors are (or were) required for a specific environment, namely evolution in a social community of peers - so perhaps an agent can be scarily effective without them, especially if their environment is very different.
If you're not careful about it though, people tend to slide into using specific definitions of "intelligence" like "calculation ability", implying it'll increase linearly as you add more GPUs to your data center. And you end up with assumptions that they'll necessarily be cold, calculating and perfect like a computer, but boredom could be a necessary property of a general intelligence, otherwise it'd get stuck on problems.
Or not. If not, maybe humans aren't generally intelligent.