Hacker News Reader

You know... this thesis that they've raised 1.1BUSD so it /must/ be good just doesn't hold water for me.

It's almost an argument from authority, and it's not very useful. In particular, investors have made large investments in the past which it turns out did not have a strong technical foundation. So there's track record here.

Yes, there's obviously something. There's obviously a team of people doing something. That's not the argument, the argument is that a 1.1B USD investment isn't really a reasonable thing to do (and that it harms the engineering environment as I noted below).

While I think the money raised is a vote of confidence from the investors -- and it's only a vote -- my point about "too big to fail" is actually more a sunk-cost argument. With that much funding, the worst-case glide path for Magic Leap might be to stumble along, sucking in continuation funding until it gets a product to market. (It's no Theranos.)

To be clear, I'm not making predictions about ML specifically. I don't have a clue about what they're doing. My general point is that huge sums of money have both momentum and gravity of their own. (I'm not saying this is rational, good, or just -- just that it seems to be true.)

Right, my comment wasn't on the last statement in your comment ("they're too big to fail"), but the bulk of it motivated by "Google didn't put ~$500m, and Alibaba something like $800m, just on a whim".

Yea, they might well use the money for /something/ useful. The might use it to acquire a bunch of other companies.

The question is, is this 1.4B USD more or less useful than 500K USD put toward the right idea.

Right. I thought it was significant that the two lead investors are strategic (the first big tranche came from Google, not GV), so there's even less signaling there.

As for $1.4B vs $500K... they're different species of investment animal. The people making the hundred-million dollar decisions aren't prepared to make hundred-thousand dollar decisions. So I don't think those huge investments are actually replacing or crowding out seed funding (especially if a highly-funded company like ML can then buy a bunch of seed-funded companies, returning liquidity to the seed market).

The only thing that's weird about it is money is fungible, so the two are truly equivalent... just not commutable. Could those billions have been a shower of seed investments instead? I dunno; this is one of those things that makes me scratch my head and shrug. I don't know enough about the history of finance, and venture finance in particular, to weigh in on the logic (or forces of nature, or chance) that led to this structure, but it seems to make sense except when it doesn't. (/hedge)

Well, perhaps think of it another way. Those billions are mostly going toward staffing costs (I would guess? If not direct employees then via semiconductor fabs etc).

Would that engineer time be better spent proving out a bunch of smaller speculative ideas rather than throwing it all behind one speculative idea.

That's the crux of it really. We could be spending our engineering time better. Maybe not funding 1000 smaller projects, but even 100 might be better.

I understand that the current investment structure probably wont support that, but it seems unfortunate.

Particularly when the large projects that engineering time gets thrown at are large, very speculative projects with likely poor DD like Magic Leap (or whoever else).

Interesting because this is the same thing I've seen having worked for my first startup. They don't want much money, and want to keep it lean, yet that for some reason turns investors away because they're used to putting in millions and millions.

How do you keep the size small, yet still get the funding you desperately need?

How often do institutional investors fall prey to the sunk cost fallacy? If you don't meet your milestones, you're probably not getting more money from the same investors. VC's make a lot of mistakes, but throwing good money after bad seems pretty rare to me.

When developing a new basic technology the milestones are often vague, and they're generally not open to DD. So hitting milestones in taken on trust and open to interpretation quite often.

Sink cost doesn't work quite the same way. It's more that they evaluate the investment as good because someone else put in a bunch of money (and they expect know what they're doing/did good DD).

I can agree that many investors will assume that because other firms invested, that is a "signal" that they use to invest themselves, and in that sense raising money can give you momentum. But I disagree that that is the same thing as the sunk cost fallacy. What I am claiming is that investors are not likely to reinvest once they realize the company is not performing as expected. It's only if they've spent their own money, and then put more money in a subsequent round that they are prey to the sunk cost fallacy.

Right, I think we agree. It's not a sunk cost fallacy. It's more sink cost signaling or something.

Facebook bought Oculus for 2 billion, and what were they buying there? Oculus had a prototype cobbled together using off the shelf parts by a teenager in his garage, a successful kickstarter campaign, and John Carmack jumped on board.

And that made Oculus worth two thousand million dollars to Zuckerberg.

The acquisition sent a strong signal, though. Shortly afterwards a bevy of competitors, big and small, started throwing out offerings.

Maybe signalling is the only reason to make these outrageous investments. Google thinks AR is important, a massive bid on an early frontrunner can kickstart an arms race which generates talent and knowledge, and even if ML fails, so long as AR in general succeeds, a consumer tech company such as Google will be able to benefit from it.

> The acquisition sent a strong signal, though. Shortly afterwards a bevy of competitors, big and small, started throwing out offerings.

In the case of the HTC Vive, given the timings of product development cycles and marketing campaigns, that was just a coincidence.

They had a lot more than that at the time of acquisition.

Huh, at first I thought ML means Machine Learning.

It sounds like Magic Leap has already given up on marketing whatever it is that they have been demoing. The Information reporter who managed to see the demo without signing the NDA said...

"The demos that I tried had some good attributes and some bad. The good - they were high resolution and they were able to focus the images at any depth - but only one focal depth at a time - so that they appeared very clear even when they were very close up. That said, the demos used a completely different type of technology from what Magic Leap ultimately wants to build. "

"Oh, and as I mentioned in the article, the bad was that the objects didn't stick very well. They were blurry and jittery when I moved my head, despite the fact that the device was connected to a desktop computer. That was one of the most difficult parts of building the HoloLens, but Magic Leap said it would figure that out on its final device. "

https://www.reddit.com/r/magicleap/comments/5hgtll/im_reed_a...

I agree that Google and Alibaba pouring on hundreds of millions adds credibility. But my skepticism is more banal: I've never seen a successful technology startup built this way.

Startups need to be nimble. When you raise 1.4 billion from the outset and hire hundreds of employees, you quickly build an internal bureaucracy. That's a hard ship to steer.

Most startups need to make course corrections, both big and small, before they find product/market fit. The political & psychological difficulty of those corrections becomes quite high when you have 100's of employees.

Investment dollars are like gasoline. Too much too early and you can actually drown out the flame.

Your faith in Google and Alibaba's judgment is weird to me. Google squandered a lot more than $500 million on terrible projects in the past (Glass and Plus for example.) I see no reason to think Alibaba is run by infallible geniuses either.

They may indeed have had a very slick demo. That doesn't mean the tech is anything special. That story has happened many times.

There's also no such thing as too big to fail.

People made all the same points you're making about Theranos.

You're right that I'm assuming there's a thorough DD process associated with those investments. It's possible ML just showed them smoke and mirrors, too. But I don't think so (partially because I've spoken with people who've seen the demo -- they're under airtight NDA, but they're clearly not skeptical of the company).

After some introspection about why I think this way, perhaps it's the sum of money that makes it seem less risky... and but it could be an artifact of scale: most investments of that size are growth capital or PE. Instead, we're seeing risk capital assigned to a potentially enormous market, so it's scaled up, too. Almost like the world's biggest seed funding? (Take that, Clinkle!)

I gotta raise a whole big stink about "but so-and-so said it was awesome and they can't be wrong".

I've been hearing the same things about the Meta 2 (https://www.crunchbase.com/organization/meta-view#/entity) for months now: https://www.youtube.com/watch?v=XxwrXacMe6Y

And then I got to try it two days ago. And it was decidedly the worst AR experience I've ever had.

I have a laser-cut box of acrylic from a Kickstarter here called the Seebright Ripple that was basically "Google Cardboard for AR" that I'm about to throw away and even that was a better experience.

You can't trust first-time user reactions to mixed-reality technologies. They get lost in the concept that they are seeing any image at all. They don't know the competitor devices on the market, and don't have any comparison to be made to properly evaluate the object.

In the Meta 2's case, the only thing they had over the Hololens was wider FOV. In every other aspect, it was worse: heavier, tethered to a computer, not adjustable enough to make the image comfortably visible, extremely bad tracking, extremely bad hand gesture input. And yet, Robert Scoble thinks this is the most important product demo he's ever seen. They could have gone out onto the conference floor at where we were and found better tracking running on an iPad [0]. It was embarrassingly bad.

So what is the Magic Leap? We won't know until they actually show us. There hasn't been any real evidence on it yet.

[0] And at this point, if you're going to be tethered to a PC, you might as well just use Vive's Lighthouse tracking. Save yourself the time, money, and bad customer reactions.

Friendly pedantic usage note: you mean by means of or some synonym (via perhaps?), not vis-à-vis. The latter literally means “face to face”, but can more broadly be used to mean something like “compared to” or “in relation to”. Your use here is unidiomatic/ungrammatical.

You're right, thanks! I'm usually the one with the grammar pedantries, but I have a lot of blind spots. I'll leave it uncorrected as a warning to the masses...

> Also, with ~$1.4 bn raised, they're too big to fail.

So many companies have raised hundreds of millions/billions of dollars and then either completely blew up or became a shell of their former selves. Nothing is too big to fail, and that mentality is toxic.

Theranos which had huge investments turned out be mmm a sad story. Big money from big companies does not mean it is true. They also fell for the hype.

Rony Abovitz must be a great salesman.