I looked at that article: https://www.researchgate.net/publica..._mould_tooling

and it didn't exactly convince me that direct injection into SLA molds was the way to go. First off, they used a SLA machine that could build parts in epoxy, which is a rare thing; most can't. Second, they just printed a shell, backed it up with aluminum-filled epoxy to draw off heat, and inserted it into a conventional steel mold. Thirdly, it didn't seem to work all that well - "flexural stress" was cited as the major cause of failure. Given that the test parts they used were extremely simple in shape, one would expect more complex ones (that would make more sense to use this method on) to exhibit even worse behavior. Even if you could make this work to some degree, without a major improvement in SLA materials I'm not seeing it as a viable process in industry.

Andrew Werby
www.computersculpture.com