While I agree that the mistake Mr. Sharp made -- it sounds like he had a network disconnection which briefly caused him to perform actions via his home IP address, rather than his VPN address -- we also don't know everything here. It doesn't sound like the guy was all that sophisticated. Using a VPN provider, in the first place, can make you a whole lot easier to be caught depending on the circumstances/provider trustworthiness/jurisdictions. I recall that there were providers which accepted cryptocurrency, but chances are good if he couldn't figure out how to block all traffic when the VPN was down, he'd have made several mistakes trying to keep the Bitcoin/Ethereum from being traced back to him.
For a crime like this -- as serious as this was, with the damages involved, the company and its internal resources/practices -- he probably had no prayer of getting away with it and in a Dunning-Kruger-like manner, he not only didn't know what he didn't know, I don't think there's any way he could have known enough about his adversary's capabilities to get away with it long term.
If a criminal wishes to be successful in getting away with a serious crime without getting caught over their lifetime, that criminal must successfully thwart detection from all current and future technologies. I mention serious because those crimes often do not have a statute of limitations these days. I'm assuming a perfect law enforcement body that similarly makes no mistakes, so a "luck factor" weighs in, but given a (not too) high-profile crime with motivation, budget, competent investigators and expanding technology, I'll law enforcement is gong to rank higher in the luck category.
It's not enough to look at what they're capable of currently. Consider this scenario: A murderer with Type O+ blood (with other common properties) strangles a man with a wire in 1980 leaving behind only that wire as evidence. In the struggle, the wire also cut the murderers hand and deposited a tiny drop of their blood on it. Being that it was a small item stored for an open case and was well preserved, it's still there, today. Luck. Back in 1980, it was of little evidentiary value. Today, that drop has a good chance of producing a DNA profile. Has the murderer been arrested (not convicted) of a felony in the last few decades? They'll probably be caught. Did a family member use certain (do they all do this?) consumer DNA services? Their family might be found, which will narrow the suspect down to a pool of people. Forget drawing suspicions by getting warrants, because it takes so little biological material and you deposit it everywhere you go, the police just wait for garbage day or follow you around town, grab something that came into contact with your mouth and they've get a profile (which will be used to get an easy warrant for a blood sample to confirm it).
Budding criminals, are you storing all of your secret plans on your drive in a bullet-proof encrypted manner and ensuring that it is airgapped? Are you doing all of your secret research on a similarly configured device, but configured to ensure all networking only works via Tor? Are you sure you didn't make a mistake that couldn't rise to the standards required to get a warrant to image your drive/take your equipment (that's hopefully turned off)? That bullet-proof encryption is rotting, and 30 years from now could represent a small hurdle above plain text.
And what happens when the time required to investigate crimes is reduced further? "We'll get around to bike theft when we're done solving all of the murders."
But what if solving a small percentage of the bike thefts went from "complaint" to "likely suspect" almost instantly if certain circumstances were right. For instance, imagine law enforcement could automate geo-fence style warrant requests (requests to get "people in a location at a certain time" from Gooble/Apple/mobile phone provider histories[0]) for every bike theft where the bike was stolen from an area infrequently traveled where and the time of the theft is known to within an hour. For any where the there was exactly one person logged, you have a person of interest -- probably the thief. Not enough evidence to prove a crime, but enough to scare some of the petty thieves into giving up more evidence through questioning (or maybe just give up). It's a stretch, on purpose -- but as technology make solving crimes less costly, less serious crimes will be prosecuted more frequently/reliably.
Full disclosure: My only credentials in this area are working in Corporate Security at a multi-national (large) telecom company for a brief stint and in a security/development capacity for most of my career; except for that brief stint, all of my work has been on the defensive/strategic side, not on the investigative side, and never with violent crimes of any kind. I simply enjoy security topics, in general, but if I've shown my ignorance in a few areas, my apologies and feel free to correct.
[0] Assuming this data is kept long enough; I am going to hazard a guess that it is a lot longer than most people think.