They pretty clearly state their metric for performance is "unicorns per $1bn" (3.08 vs 1.22).
They're suggesting dollars invested in UK startups are more likely to create a unicorn than those same dollars put into us startups, hence higher performance.
Quite typically the highest return investments are picked first; as such there is a diseconomy of scale, it should be of no surprise that the ROI decreases as investment scales. Perhaps instead of thinking of it as a sign of efficiency it should be thought of as an underdeveloped market with the UK foolishly leaving money on the table.
Also it is not measuring average returns per $ but returns split by company then a vector operation of ceiling (1bn) divided by my 1bn then aggregated.
Thats a dumb metric. It is like saying conditioned on a doctor saw you, how many people were saved. You gotta count folks who died because they did not see the doctor too.
Some of the unicorns could be simply rent seekers making money by colluding with government people. Investors saw them sure shot and hence invested. That model is neither scalable not healthy.
Well, yeah. Presumably the way they achieve a good ratio is by being more discerning with their investments. If you try to scale that you get the US's numbers as you throw money at anyone with a pitch deck.
The thing is, the US model is still better in aggregate as it basically results in every single unicorn being birthed (in theory, there’s simply no lack of funding for plausibly good ideas). Focusing on a conversion metric is a vanity exercise when absolute value creation is available.
We would all rather own a huge low margin business than a tiny high margin business. Eg, 5% margin on $1B of sales is better than 80% margin on $10M sales.
I know nothing about UK investment culture, but I have been building startups for 15 years, volunteered for multiple accelerators, made angel investments and rubbed shoulders with many US VCs.
My impression of the US VC industry, especially near the end of the zirp era and even more especially during covid, the amount of VC capital being deployed far surpassed the number of competent VCs or startup founders.
Many VCs were making decisions based on factors that were NOT correlated with future venture success. Many VCs in fact probably biased companies away from future success, because they didn't understand what they were doing and their instincts from prior industries or investment regimes were directly the opposite of what was needed for an early stage startup.
In other words, there is risk aversion, and there is foolishness. I know for a fact there was a lot of foolishness going on in us VC investing (I even participated in some of it, learned from it, and I know better now).
I'm not just talking about VCs throwing money at anyone with a pitch deck. I'm talking about VCs having a backwards understanding of what makes startups successful and actively pressuring startups that could have worked into doing the wrong things.
There is a core of US VCs that are the world leaders in what they do, exceptionally aware of what makes startups successful and have the track record to prove it - this minority is overwhelming responsible for the industry's ROI. There is also a massive graveyard of fools who tried to replicate that success and failed for a variety of reasons.
Also if the UK investor says "we'd love to invest, can do the whole round, want you to send us your current cap table, we need to get the deal done before April because we have capital to deploy before end of the current financial year" don't actually expect the funds.
I had the misfortune to be pitching to VCs in the UK, the money was so expensive that it’s just not worth it, much better off moving to the US if you can.
That might be a case of selection bias. The only reason to put up with UK regulations and other issues is if you have a compelling reason to do so (compared to similar opportunities).
But a per capita figure is diluted by how big of a tech sector that country/region has relative to the rest of its economy. The more people work outside of that sector, the lower is any per-capita figure from tech. Fewer lines of code written per capita, fewer bugs per capita, ...
The figure per invested dollar is much better: how many unicorns emerge per billion of investment money.
Those who chase unicorns are mainly investors (plus people who want to join startups that become unicorns). That figure is directly relevant to them.
The tech sector in the UK is pretty big but also massively finance weighted. And salaries a lot lower than the US so a $ investment goes further. At my company a few years ago our one developer in Boulder earnt more than our head of software in the U.K.
I'm in the middle of doing a major version upgrade of postgres from pg15 to pg18 using the same logical replication techniques that this article talks about. As this article mentions, dropping the indexes on the new database before replication is key. Otherwise the replication takes forever because each insert needs to update the index as well.
This is the same even if doing bulk or batched inserts. MS's process of importing an exported DB (when using that method, i.e. with Azure SQL rather than an on-prem restore to on-prem where a page-level backup is of course much more efficient) doesn't create indexes until after the full data import has completed.
It is being restructured to no longer be a nonprofit, but a for-benefit corporation instead. That is why California's approval was required (and IMO is just as corrupt as it sounds).
"OpenAI and Microsoft have made their next move in their attempt to expropriate the OpenAI nonprofit and pull off one of the largest thefts in human history."
Corporate law is overwhelmingly state law. Every federally tax exempt entity is a state (or foreign) corporation or other kind of entity, and states (or foreign governments) impose rules on corporations registered in their borders.
Plus, many states levy their own corporate taxes. A nonprofit corporation needs to secure tax-exempt status from states as well as the federal government. This is a necessary implication of America's dual-sovereignty system.
> aren't non profits a federal thing with rules dictated by the IRS? Why is California involved?
Most states incorporate federal rules for their own exemptions for charities and non-profits. California treating OpenAI to date as a non-profit has revenue implications for Sacramento.
Corporations and nonprofits are a state regulated thing. The IRS only gets involved to approve whether donations are federally tax deductible and things like that, which apply to federal tax laws only.
The Usa is divided into states, each with their own legal jurisdictions. Businesses tend to pay taxes to their state. OpenAI is headquartered in San Francisco, California.
The more interesting question is will they be able to deduct those past losses (while they were "nonprofit") from their future income in the same way they could if they had been a normal corporation all along.
you can IPO the for-profit subsidiary. done before. the main tricky part is resolving the tax issues since as a non profit you are exempt, but obviously a for profit is not exempt from paying taxes
Mozilla has this same structure (non-profit parent, for-profit subsidiary) because the subsidiary has to pay taxes on the money Google pays to be Firefox's default search engine.
As do many gift shops attached to non-profit museums and art galleries.
The difference is that Mozilla’s for-profit arm is owned by the non-profit. The for-profit part of OpenAI is there to make a lot of individuals and for-profit corporations rich.
OpenAI’s profit-generating subsidiary isn’t just there to further the non-profit mission like a museum gift shop or Mozilla’s for-profit subsidiary.
That’s why this was scrutinized more and the California AG got involved but it’s not that rare either.
Novo Nordisk the maker of Ozempic for example IPOd, diluting the Novo Nordisk Foundation’s share (though they still have controlling voting rights due to share classes IIRC) to raise money. SRI International spinoffs often get sold (Siri) or raise money and IPO (Nuance) diluting the nonprofit’s share significantly in the process.
A nonprofit that owns a for profit subsidiary is no different than a regular shareholder and can decide that diluting to reward employees or get investors is worth it to grow the value of the whole company.
There's so, so much misinformation about this out there.
For example - every US nonprofit starts as a plain old vanilla C corporation, and then applies for 501(c)3 status which the IRS may or may not grant. It's a privilege to be a nonprofit.
The punishment that may be levied on a nonprofit is ... loss of that status and a return to a commercial corporation. That loss of status might have knock-on impacts on things like, say, tax deductions offered to donors, and I guess possibly on corporate income tax to the extent a company's accounting shows a profit. But it's not a thing you're "locked into" somehow and trying to escape. Quite the opposite; it's a thing the Federal government chooses to support financially as a matter of public policy.
oAI had a lot of work to do to get recapitalized like it did, but it was not the non-profit status that was the (major) problem. It was (at the least) the investment covenants made with the Microsofts of the world that bound them; the MS deal was the big thing here.
I think it’s generally more accurate to say that charitable assets are likely either locked to charitable purposes or need a fair valuation in case of disposal or wind down. I don’t know all the details here, but I would guess it’s enough of a mess between the parent and the for profit sub that some negotiation was inevitable.
I'm literally in the middle of upgrading my prod db to pg18. Its about 6tb, has a few thousand queries per second, should I be considering running in 'worker' mode instead of 'io_uring'?
For upgrades which have enough risks as it is, I would keep the number of variables low. Once upgraded and stable, you can replicate to a secondary instance with io_method switched and test on it before switching over.
MS was sizing them up a short time ago, I would imagine it'd be something strange like laying everyone off then hiring them again, or moving the IP to a child corporation Firefox-style
Because when you send a document for a $75MM contract to be signed you want to send it via a well known and trusted document signing platform like Docusign where you know everything is legit and legally defensible.
I'm currently self hosting my notes/journal/knowledge base with Trilium, photos with Immich, and files with File Browser, very happy with that setup so far. I just like the feeling of knowing I own my important data and that it won't go away because some third party company goes out of business or sunsets an app.
