Procrastination accompanied by guilt comes from an internal conflict about whether one should do the activity. Sometimes the conflict comes from partly wanting to cater to one's present self (by engaging in more gratifying short-term activities) and partly wanting to cater to one's future self, or to others (by doing something that's less rewarding in the short term but that will pay for others, or pay off in the long run).

But this is often not the only element present when one procrastinates and is guilty. Often another element present is uncertainty as to whether the activity is what one should be doing, even when considering the indirect consequences. This can be subconscious: one might consciously think "I know I should be doing X, but I just can't motivate myself to do it" while simultaneously believing on some level that one shouldn't be doing X (even when considering indirect consequences). The conscious self isn't always right in these situations – sometimes rather than trying to overcome procrastination, one should instead abandon the activity that one is procrastinating, for example, when the activity is neither interesting nor important. The subconscious self isn't always right either: sometimes it's operating based in false information or insufficient reflection.

If one can recognize and resolve the uncertainty, this can increase one's motivation to do the work if it's the right thing to be doing, and help one decide not to do the work if it's not the right thing to be doing. So determining whether there's uncertainty and trying to resolve it can have high value.

It's not always possible to resolve the uncertainty. When this is the case, recognizing that there is uncertainty may not be helpful. Unfortunately, uncertainty can be demotivating even when completing the task is expected value maximizing. The question of how to stay motivated in the face of uncertainty is an important one that I don't know the answer to in general.

Below, I give some examples of beliefs that can coexist with "I know that I should do the work" that give rise to uncertainty, together with commentary. Some of the beliefs described overlap in character, or can be present simultaneously.

A belief that it's more effective to do the work later on

Sometimes there are higher priority things to do (even if one should do the work later on). Sometimes one is in an unusually poor state to do the work (for example, if one is sleep-deprived and this is not a regular condition). In such cases, procrastination can be rational.

Sometimes one rationalizes procrastination with the justification that there are higher priority things to do in the near term, even when it's not true. Sometimes thinking that one will be in a better state to do the work later on is wishful thinking. So this belief may or may not be good reason to abandon the activity.

A belief that one can't do the work

Sometimes the belief is well-grounded, for example, for most people who are working on solving a famous unsolved mathematical problem or working creating a tech startup. It tends not to be true for people who are trying to do things that many others have done successfully before. Sometimes the belief can arise from it not immediately being clear how to do the work, even though one could figure out how to do the work if one thought about it. For example, if one is having trouble learning to code, one can ask friends for help, or use Google to find answers to questions.

A belief that one is poorly suited to the work

Even if one can do the work, one might procrastinate it because one has the sense that even if one does it, it won't move one forward.

I know a number of former engineering majors who found it very hard to motivate themselves to work on their first year math, science and engineering classes because they struggled to learn the material, decided that engineering wasn't for them, and felt liberated upon coming to this conclusion, feeling much better doing work that they're better at and enjoy more.

Their motivational problems may have been a valuable signal to their conscious selves that they should be doing something else, and their decision to drop engineering may be rational: they could have been picking up on not being good enough at engineering (or find it enjoyable enough) to be able to get good engineering jobs relative to the other jobs that they could get.

They may have underestimated their ability to improve (c.f. How my math skills improved dramatically). They may have been misinformed about the extent to which engineering jobs are similar to learning the material in the required courses. So their procrastination may not have been a reliable signal that they should abandon the path that they were on.

A belief that the work is inefficient or unimportant

All else being equal, we flinch away from work that's inefficient or unimportant. So procrastination can be a signal of this belief.

Sometimes there's a better way to accomplish the goal that a task is supposed to accomplish. For example, it might be possible to write a computer program that automatically carries out a tedious task, whether it be computational, information-gathering or sorting. Also, carrying out the task may not help achieve the goal at all. 

On the flip side:

• Even if it's possible for a goal to be achieved more efficiently in the abstract, one may not have the resources to accomplish it more efficiently.
• In the modern economy, many important jobs are several steps removed from the tangible results, so that one can get a subjective sense of not getting anything important done even when one is.
• Even if work is of no intrinsic importance, it may still be important to do it so as to meet credentialing requirements (e.g. in the context of school).

So here too, it may or may not be rational to act on this belief.