The Chinese language needs to have spaces. It would make learning & reading faster, at very little cost.
They would not have to be large spaces - just a tiny space, 1/4 as wide as a character.
In ancient, grid prints, this would not work - but today, most text doesn't line up anyway, so adding spaces wouldn't make it look any worse.
It would be easy to put spaces as you were writing - most input methods can already detect words (since there is so much help going on during writing, anyway), Plus, the person writing the sentence always knows where words end and begin.
Autotranslation would be a lot easier.
Reading would be easier for people whose chinese isn't perfect. This would be good for foreigners, and also good for native speakers whose reading level isn't very high. Helping people like that is a supposed goal of the chinese government.
Most controversially, I think that reading would be faster for everyone, at every reading level. This is because even fluent readers are now doing work that spaces would eliminate. Having to identify word boundaries is actual work - it takes CPU time to do on a computer. Your brain doesn't just do things "for free" - although it does have specialized hardware, running it still takes energy.
The biggest one is defensiveness - people who can read Chinese have put so much effort into it that they don't want to make it any easier for an outsider.
People also are in denial that it would matter. But, people are always in denial about how changing things could ever improve them. This paper: http://research.chtsai.org/dissertation/ suggests that most ambiguities can be resolved pretty easily - but I don't think that's sufficient to say spaces don't matter, because adding spaces would solve all of them, not just most. He mentions a figure of 93% of ambiguities being trivial to resolve - which seems like a lot, but considering thousands of words per hour read, and the increased cost of the failure case, it could be a significant cost.
In chapter 4 of the previous dissertation there is mention of research favoring both sides, but without details about how it was conducted, the claim that spaces would not improve reading speed has not been resolved.
Many languages were originally written without vowels at all. Ideographic language such as Japanese, Egyptian, and Mayan also originally had no pronunciation help - but over time these were added. Languages added dots and lines above consonants to show vowels, and eventually added vowels as actual letters. Japanese added furigana (phonetic help characters) that can be used when necessary to show pronunciation, and as a phonetic alphabet. Other ideographic languages added phonetic shorthands, too.
In all of those cases, imagine what it would have been like to try to argue before the change, compared with after. Before the change, Hebrew scribes probably venerated the ability to read text quickly, without vowels or any help. Learning that skill was considered to be a valuable spiritual process, to contain inherent value and give mystical power and wisdom. And they probably resisted the change. But now that the change has been made, nobody wants to go back.
These same people don't advocate making the language even harder, removing or conflating certain letters, or adding even more special cases that need to be memorized. If having hard languages is good, we should make them even harder! There would be plenty of ways to do it. Making the claim that we are already at the ideal level of difficulty is unreasonable, since if you look at the actual reasons languages changed, you will not find much freedom of choice.
Where we are today in language evolution is due to the history of colonization, physical division of cultures, and communication frequency between cultures - not due to rational choices about language complexity and simplification.