“Corporate VC” is an oxymoron. It’s like a division of a big company pretending they can operate “like a startup.” It’s like the lie that a company won’t change once it has been acquired. It’s like a cat pretending it’s a dog.

This is true for Valley-based technology companies who used to be startups and now try to invest in the next generation. Imagine how a global multinational corporation is going to fare. Spoiler: Not well.

DealBook wrote today about big multinationals opening up around Palo Alto to fund startups, and GigaOm followed up with a report on international companies upping their VC game too. I think I wrote this exact post a decade ago. Unfortunately, not everything was archived on the Web back then, or I’d link to it. That time around, corporations’ optimism that innovation could magically rub off on them and shaky startups’ hopes for a lucrative channel deal were mostly all dashed.

There’s a reason that so many of these programs were started in the last two years: Because hundreds of them were shuttered when they didn’t work in the last cycle.

Here’s DealBook’s quote from the guy from American Express — whose corporate VC office is hilariously located in the old Facebook headquarters. They literally hope innovation rubs off on them:

“American Express is a 162-year-old company, and this is a moment of transformation,” said Harshul Sanghi, a managing partner at American Express Ventures, the venture capital arm of the financial company. “We’re here to be a part of the fabric of innovation.”

I have never once heard someone actually innovating say a phrase like “fabric of innovation.” A 162-year-old company trying to learn from startups is like your grandpa picking up skateboarding from the kids down the street. Best case scenario, it’s going to lead to a broken hip.

The article continues to talk about how these companies have an edge over VCs because they care about the financial return, they just want to see what innovation is.

Forgive the snark, but I’ve seen this movie before. There is only one company that has consistently — cycle after cycle –maintained a top corporate venture capital program: Intel Capital. In the wake of the last bust, it was one of the most prolific investors at a time when many traditional firms were burned, and almost all corporate investors were writing off failed deals and slowly backing away. BlueRun Ventures– which started life as Nokia’s corporate venture arm– has had a decent run as well, although its biggest win, PayPal, was a while ago.

And there’s one thing that has saved Intel Capital: It does the opposite of what most corporate VCs do. It invests for financial return first, and strategic fit second. Google Ventures copied this strategy, and Google has one-upped Intel by hiring big name partners who get deal flow like Joe Kraus and, more recently, Kevin Rose.

But these are very rare exceptions. Most corporate VCs live in an awkward parallel universe where they are both a bad fit in-house and a bad fit for the startups they fund.

Let’s deal with in-house first: Companies can’t possibly compensate corporate VCs like traditional partners. If they did, they’d make more than a lot of executives who do more centrally important jobs for the companies.

It’s a very real problem that means hiring and retaining people who actually have some insight into building companies is near impossible, especially if the job requires the patience to navigate a huge corporate bureaucracy at the same time. At best, corporate VCs are a bit like minor league baseball teams. If you’ve got the chops to be a real VC, you’re either moving up or moving down. Either way, at many companies it’s a revolving door of talent and climbers.

And that’s one reason why corporate venture programs tend to come and go within big corporations — their advocates do too. And, making matters even worse, they frequently come into vogue at the worst times in the venture cycle — when valuations are at their peak and about to decline. This was certainly the case in the early 2000s, and one could argue it’s the case now.

Big companies are just fundamentally at odds with startups. For these deals to work, there has to be an uncanny strategic fit. There needs to be something those two companies specifically need from each other right now — not at some vague time in the future.

Microsoft’s $15 billion valuation investment in Facebook was one of those. Microsoft was just desperate to be associated with Facebook in any way, so much so that it outrageously overpaid. And the deal came with a deal to resell its ad inventory — a counter against Google’s similar deal with MySpace. And what did Facebook get? Well, Facebook wanted money without giving up much equity. This was the Yuri Millner deal before Yuri Millner invaded the Valley.

An even better example — seemingly — is the recent Starbucks/Square deal. Starbucks was hungry for a partnership with a mobile wallet player, and Square can dramatically benefit from the massive store footprint. Also, Starbucks’ chief executive Howard Schultz is one of those rare entrepreneurial guys still running a big company.

In both cases, each party had a specific thing they needed — and no expectations beyond that.

But in the majority of cases, corporate VCs are hoping to get an option to buy a company in the future. And if an entrepreneur does a deal based on that, he’s sort of implicitly saying he doesn’t think his company is going to make it. With more unconnected corporate (read: dumb) money rushing at startups, I expect there to be adverse selection: The worst startups will largely be the ones taking the cash.