As one lesson from this debacle, I think it's useful to ask ourselves what are other areas of groupthink that leave us in the tech industry particularly vulnerable?
E.g., as an OSS developer, I've used the MIT license for 10+ years because other devs I trust use it, but I don't have any real reason why.
What are other examples to be aware of?
The fact that we're less than 24hrs from the events that just happened. Where the banking system was just put at risk by the actions of VC (including many prominent in this community) and that it's *not* the top post on this site is a very clear red flag. Really a "tiny c compiler" is whats most pressing right now?
As well the fact that any post I've seen from made this morning critical of the actions people in the community took over this weekend have all been flagged off the front page is another clear red flag.
I won't link it here, but I feel like that Stratechery article is must read at this point and clearly captures a big part of the problem.
This currently has made it up to #4 - I'm willing to be admit if I'm proven wrong. But let's see how long this one lasts.
----
Edit: and no more than 3 mins later after I posted the above and this this reaching the top of HN it was immediately dropped down to ranking #88. In the span of 3 mins.
You want to know what group think there is. *That* right there is the group think you should be worried about. People aren't even comfortable with being criticized about what happened this weekend.
* Executive-level hiring. How many startup founders bring entire crews of executives that are recommended by VCs. Those transplanted execs then bring in their own "crew", and proceed to not only change the culture, but also repeat many of the same mistakes, pursue the same processes, repeat the same groupthink that has been done before. It reminds me of the exec hired in HBO's Silicon Valley who developed the "box" and brought in those sales and VPs.
* Processes. How many startup founders blindly adopt OKRs, growth strategies, and methods without even thinking if they are relevant to their industry, business, culture, or skills?
* Technologies. VCs push (forcefully sometimes) the use of technology by their other portfolio companies, leading to lots of misfit applications. There's too much cloud-centric thinking which causes problems. The cloud is great, but its not a one-size fits all. Some independent thinking is helpful. Not everything needs to be in the cloud
* Grow grow grow, then fire fire fire. The grow-fast then fire-fast approach mentality is really harmful.
* Following VC recommendations for core business operations, such as who to hire for law firms, accounting, HR, and other operations. Is it a wonder that so many startups make the same mistakes, do the same things, and run the same processes when they are all using the same vendors? Not too long ago, banking with SVB was "just a thing we all did". Look where that group think gets you. Maybe you don't need that law firm, accounting firm, or HR just because every other startup you know is using them.
* Idolization. There's too much worship of enterpreneurs, investors, etc. Yeah some of these folks have done some great things. But don't take every word they say as if it's gospel. It's ok to disagree with PG, Elon Musk, Peter Thiel, YC, etc. etc. whomever you idolize. They're not always right, and perhaps they're often wrong. Groupthink often happens when people exhibit cult-like behavior by following the herd or guru.
* Diversify your sources of information. Don't read so much HN or TechCrunch or whatever without balancing that information diet with other sources that aren't part of the same sphere of thinking. Groupthink happens in echo chambers and information bubbles. HN is one.
Hypervisor vulnerabilities. It's fundamentally risky to have processes with different security contexts run on the same physical hardware. It's probably just a matter of time before we see some really bad hypervisor jailbreaks. All cloud computing is at risk.
Mass surveillance through smart devices. The cameras and microphones are already in all of our our houses. Just a matter of time before somebody turns them on.
If you go back in US banking regulation history, there used to be very strict rules around branching (which is one of the reasons why there is such a large number of regional banks in the US). This had the side-effect of preventing banks from having a diversified set of depositors, so local shocks (crop issues, weather, etc.) were much more likely to decimate small, local banks, which is one of the reasons why bank failures used to be MUCH more common.
The parallels between cloud and OS are wide, as many basic functions that use to be done on host systems are commonly done in cloud. Database, message buses, application state, processing, application logging, file storage, cacheing, user authentication. In twenty to thirty years many of the cloud software being built today will still be running, and still be generating revenue but will feel stifled and limited in the same way working on a 25 year old Windows-only application would feel today.
I'd love to see "a cloud in a box", or a collection of open source tools to provide a full range of cloud services powered by open source software that can be deployed on arbitrary VMs and sanely configured out of the box to interact with itself. This sort of grouped redistribution of preconfigured software paired with an installer could look to Linux distros as an example from the previous generation. This would remove vendor lock in, and insure any critical tools can be maintained by the companies that relay on them.
- How many women do you have on the team?
- Well, not many
- But how much?
- None
* Looking for a "culture fit"/"perfect fit" while hiring. Surprisingly realized that as immigrant, I have better interview performance while interviewing with other immigrants. I had Facebook offer because other folks were from China, India, and so on. However, white males almost always look for a "fit". If you can't make friends with them in one hour, in most cases you gonna hurt their "subjective feeling". Many many times I was getting response like "everything was good, but it's not just a good time". Hello Roblox, Gopuff, Coinbase, etc. Interviewing with immigrants is such a relief, like it or not!
There are more guys in VC named Andy or, to choose a totally random name, Peter, than there are non-guys, and they know who they know.
The point isn't to shame people in the system, but just like diversifying investments is good not for show but because it actually works, we need better tech elites who have a way way way broader understanding of the world and the only way that's possible is to have people in this world from a far wider range of backgrounds.
SVB collapsing was far less likely than AWS going down due to a cyberattack.
It seems that one of the core issues here is that SVB was in charge of getting "yield" from their deposits on top of the value that startups themselves were trying to generate. From a finance POV it make sense that you'd want to be earning interest, and trying to keep pace with inflation, but in an economic sense it also seems like double counting to expect VC capital to generate risk-free return while waiting around for the startup's businesses to pay off.
I understand that every business looks like this to some extent, but in the VC funded startup world it is a unique amount of cash invested in a lot of businesses that need stability more than they need bank-balance interest. The risk profile of SVB turned out to be highly correlated to the risk profiles of the companies they banked.
The groupthink at the moment is phew, the Fed, FDIC and treasury swooped in to save the day (and the depositors).
But, the root cause lurks.
How many other financial institutions are sitting on leveraged assets similar to SVB and might fail a slightly tougher stress test?
But lot of things that most people do are safer than the unbeaten path, even if it fails most likely there is a bigger chance of being saved. Like now, seems like the US gov. is stepping in.
But of course crypto was a huge groupthink investment, all the big investment firms & banks went in just because they couldn't ignore the years of speculation, but not because of growing usage in society leading to increased "real world" value.
I've seen huge groupthink in cloud hosting. Im not sure I really understand it, especially from within governmental organisations they've spend millions moving to the cloud with little to show for, higher monthly costs, lower delivery times.
People used SVB because (1) It was a normal, trustworthy bank with many years of regular operations, (2) It was one of the few banks that would work with early stage startups that have lots of funding, (3) in some cases they were basically forced to by the VCs that funded them and (4) SVB had a well-earned reputation (and now well-lost).
The collapse of SVB was not made by any group. It was made by a relatively small number of individuals who made bad investment decisions. These is no reason to believe that they bought MBSs because of groupthink.
Bank makes schoolboy errors when it comes to investment, but people who wanted to protect their money get the blame for its collapse.
Come on.
Whether a silicon valley startup succeeds or fails already takes large influences from a survivorship bias, and they're all basically going to cheat in some way to establish themselves. Whether that's Uber ignoring local taxi laws, Atlassian messing up multiple customer accounts, or some web service you bet your business on that has never tested their backups.
So with a decision like banking say a series A startup with their first $5million is really clever and splits their money evenly between two banks, and one bank fails. That may still be just as crippling to a startup that now needs to distract themselves with fundraising again earlier than expected, that doesn't have a proven track record, and just lost half their money that their investors may still demand a return on.
So the survivorship bias might just be that the startups that chose the other bank, as one less decision to dwell on and distract them, are the ones that survived. Or the startups that were extremely diligent and spent a great deal of time de-risking themselves from bank exposure died... because they spent too much time not getting a product to customers.
So maybe the better question is, at what point should a company to begin to assess banking as a risk to be mitigated.
I hope that this lesson should put the 'growth over profit' narrative to rest.
Seems like tech is systemically risky and should be highly regulated.
SVB made terrible investments and gambled hundreds of billions of dollars, and then they lost. It's not groupthink the pull your money out of a sinking ship, it's idiotthink to keep it there.