Author: azeeadmin

I’ll be heading back to Europe on December 18th to run a pitch-off in Wroclaw, Poland. It’s a bit out of the way but well worth a visit if only for the sausages.

The event called In-Ference is happening on December 17 and you can submit to pitch here. The team will notify you if you have been chosen to pitch. The winner will receive a table at TC Disrupt in San Francisco.

I’m also thinking about an event in Warsaw on the 21st but WeWork didn’t look doable (and I don’t like co-working spaces). If anyone has thoughts on a new venue drop me a line at john@techcrunch.com. Otherwise, I’ll see you in Wroclaw! Wesołych Świat!

Companies are staying private longer and longer and New Enterprise Associates (NEA) has a new plan to combat the negative implications of the trend. The storied venture capital firm has invested in a spinout fund called NewView Capital and sold 31 of its late-stage portfolio companies to the effort.

Former NEA general partner Ravi Viswanathan is leading the fund, which today announced a $1.35 billion pool of capital to invest in growth-stage companies across industries and support the large cohort of former NEA investments now under its purvue. Among the companies purchased by NewView from NEA are 23andMe, Acquia, Canopy, Duolingo, Forter and GumGum.

The secondary deal relieves some of the pressure on NEA to provide liquidity to the investors in its older funds, known as limited partners (LPs), without encouraging its portfolio companies to exit before they’re ready. Consumer genetics company 23andMe, for example, was founded in 2006; NEA has been among its investors since 2007, per PitchBook. It typically takes between five and eight years for a VC-backed startup to be acquired or to pursue an initial public offering, making 12-year-old 23andMe long overdue for an exit.

“There’s a lot of congestion on the road to liquidity,” NEA managing general partner Scott Sandell said in a statement. “Market dynamics are compelling companies to stay private longer, which creates growing demand for follow-on dollars and stretches investor holding periods to a decade or more.”

Sandell has led NEA since the mid-1990s. Viswanathan joined him in 2004 and has co-lead the firm’s growth equity investment practice for several years. He’s been focused on the NewView spinout since the beginning of 2018.

“We are providing liquidity to a set of companies that would require several years longer to get liquidity,” Viswanathan told TechCrunch. “It’s a block sale that has never really been done and it’s a highly curated basket … There have been GP restructurings in the past; that’s not this. There have also been lots of one-off purchases; that’s not what this is either. This is really a strategic spin-off of a curated set of companies done at scale.”

NEA, founded in 1977, is known for its investments in Snap, Jet and more recently, The Wing, Opendoor and Casper. The multi-stage investor wrapped its 16th growth fund in 2017 on $3.3 billion.

The firm has been making strategic moves over the last several years to manage changes in the VC market exacerbated by SoftBank’s nearly $100 billion Vision Fund. In 2015, for example, NEA raised its first opportunity fund in addition to a multi-billion flagship vehicle so it could make follow-on investments in its buzziest companies. Doing that gave it the flexibility to pay a bit more for the swelling late-stage valuations some of its best investments had conjured up. The firm was also one of the first to raise a “mega-fund,” or a fund larger than $1 billion, in 2000.

Since then, funds have only become larger and companies more expensive. NEA’s latest decision to separate some of its investments may be especially wise as many are predicting an imminent turn in the venture market.

Joining Viswanathan at Burlingame-based NewView as operating partners are Tim Connor, a former NEA executive-in-residence; and former SigFig chief product officer David Yoo. Prashant Gangwal, the director of finance at SharesPost, joins as chief financial officer and chief operating officer. 

NewView’s LPs include Goldman Sachs and Hamilton Lane.

Google -owned AI specialist, DeepMind, has claimed a “significant milestone” in being able to demonstrate the usefulness of artificial intelligence to help with the complex task of predicting 3D structures of proteins based solely on their genetic sequence.

Understanding protein structures is important in disease diagnosis and treatment, and could improve scientists’ understanding of the human body — as well as potentially helping to support protein design and bioengineering.

Writing in a blog post about the project to use AI to predict how proteins fold — now two years in — it writes: “The 3D models of proteins that AlphaFold [DeepMind’s AI] generates are far more accurate than any that have come before — making significant progress on one of the core challenges in biology.”

There are various scientific methods for predicting the native 3D state of protein molecules (i.e. how the protein chain folds to arrive at the native state) from residual amino acids in DNA.

But modelling the 3D structure is a highly complex task, given how many permutations there can be on account of protein folding being dependent on factors such as interactions between amino acids.

There’s even a crowdsourced game (FoldIt) that tries to leverage human intuition to predict workable protein forms.

DeepMind says its approach rests upon years of prior research in using big data to try to predict protein structures.

Specifically it’s applying deep learning approaches to genomic data.

“Fortunately, the field of genomics is quite rich in data thanks to the rapid reduction in the cost of genetic sequencing. As a result, deep learning approaches to the prediction problem that rely on genomic data have become increasingly popular in the last few years. DeepMind’s work on this problem resulted in AlphaFold, which we submitted to CASP [Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction] this year,” it writes in the blog post.

“We’re proud to be part of what the CASP organisers have called “unprecedented progress in the ability of computational methods to predict protein structure,” placing first in rankings among the teams that entered (our entry is A7D).”

“Our team focused specifically on the hard problem of modelling target shapes from scratch, without using previously solved proteins as templates. We achieved a high degree of accuracy when predicting the physical properties of a protein structure, and then used two distinct methods to construct predictions of full protein structures,” it adds.

DeepMind says the two methods it used relied on using deep neural networks trained to predict protein properties from its genetic sequence.

“The properties our networks predict are: (a) the distances between pairs of amino acids and (b) the angles between chemical bonds that connect those amino acids. The first development is an advance on commonly used techniques that estimate whether pairs of amino acids are near each other,” it explains.

“We trained a neural network to predict a separate distribution of distances between every pair of residues in a protein. These probabilities were then combined into a score that estimates how accurate a proposed protein structure is. We also trained a separate neural network that uses all distances in aggregate to estimate how close the proposed structure is to the right answer.”

It then used new methods to try to construct predictions of protein structures, searching known structures that matched its predictions.

“Our first method built on techniques commonly used in structural biology, and repeatedly replaced pieces of a protein structure with new protein fragments. We trained a generative neural network to invent new fragments, which were used to continually improve the score of the proposed protein structure,” it writes.

“The second method optimised scores through gradient descent — a mathematical technique commonly used in machine learning for making small, incremental improvements — which resulted in highly accurate structures. This technique was applied to entire protein chains rather than to pieces that must be folded separately before being assembled, reducing the complexity of the prediction process.”

DeepMind describes the results achieved thus far as “early signs of progress in protein folding” using computational methods — claiming they demonstrate “the utility of AI for scientific discovery”.

Though it also emphasizes it’s still early days for the deep learning approach having any kind of “quantifiable impact”.

“Even though there’s a lot more work to do before we’re able to have a quantifiable impact on treating diseases, managing the environment, and more, we know the potential is enormous,” it writes. “With a dedicated team focused on delving into how machine learning can advance the world of science, we’re looking forward to seeing the many ways our technology can make a difference.”

We’re entering the final days leading up to Startup Battlefield Africa 2018, which takes place on 11 December. We have our passports, we’re packing our bags and we can’t wait to arrive in Lagos, Nigeria to watch 15 of the continent’s best startups launch to the world.

Will you be there cheering the competitors and celebrating the richness of Africa’s growing startup scene? We’ve sold out of tickets to the day-long event, but — and this is great news — we set aside a limited quantity of tickets that we’re giving away for free. If you want one, act now and apply for a free ticket before they disappear.

Center stage of this day-long, action-packed event is, of course, our renowned pitch competition, Startup Battlefield. Out of hundreds of applications, we winnowed the field to 15 remarkable early-stage startups across Africa. Startup Battlefield consists of three preliminary rounds — five teams per round — followed by a final round.

Teams have only six minutes to pitch and demo their product to a panel of expert judges. After each pitch, the judges run the presenting founders through a gauntlet of tough questions. Up to five teams will move on to the finals to pitch and answer questions again.

One team will claim the Startup Battlefield Cup, along with US$25,000 in no-equity cash. Plus, the winning founders also win a trip for two and the opportunity to compete in Startup Battlefield at a TechCrunch Disrupt in 2019.

Between rounds, we have a slew of outstanding presentations and panelists scheduled to discuss the most pressing issues facing early-stage startups in the region. Check out the conference agenda, but some of the people you’ll hear and learn from include:

• Omobola Johnson, a senior partner at TLcom Capital, is the former minister of communication technology for Nigeria. Her 25-year tenure at Accenture — as managing director — informs her expansive knowledge of startup investing.
• Lexi Novitske, the principal investment officer for Singularity Investments, manages the firm’s Africa portfolio. She’s a proponent of understanding local complexities, modifying Western business attitudes and working with companies to ensure desired results.
• Marième Diop, an investor at Orange Digital Ventures Africa, focuses on early-stage African startups, and she’ll be talking all about venture capital.
• Shikoh Gitau, the head of product at Safaricom’s Alpha incubator, will share her perspective on the talent and innovation within Africa’s expanding startup landscape. She’ll also talk about repatriating entrepreneurs.

Come and join us in Lagos on 11 December. Experience Startup Battlefield Africa 2018, engage in world-class networking and spend a full day celebrating the awesome talent of the region’s startup community. All you need to do is apply for your free ticket right here.

AWS, once a nice little side hustle for Amazon’s eCommerce business, has grown over the years into a behemoth that’s on a $27 billion run rate, one that’s still growing at around 45 percent a year. That’s a highly successful business by any measure, but as I listened to AWS executives last week at their AWS re:Invent conference in Las Vegas, I didn’t hear a group that was content to sit still and let the growth speak for itself. Instead, I heard one that wants to dominate every area of enterprise computing.

Whether it was hardware like the new Inferentia chip and Outposts, the new on-prem servers or blockchain and a base station service for satellites, if AWS saw an opportunity they were not ceding an inch to anyone.

Last year, AWS announced an astonishing 1400 new features, and word was that they are on pace to exceed that this year. They get a lot of credit for not resting on their laurels and continuing to innovate like a much smaller company, even as they own gobs of marketshare.

The feature inflation probably can’t go on forever, but for now at least they show no signs of slowing down, as the announcements came at a furious pace once again. While they will tell you that every decision they make is about meeting customer needs, it’s clear that some of these announcements were also about answering competitive pressure.

Going after competitors harder

In the past, AWS kept criticism of competitors to a minimum maybe giving a little jab to Oracle, but this year they seemed to ratchet it up. In their keynotes, AWS CEO Andy Jassy and Amazon CTO Werner Vogels continually flogged Oracle, a competitor in the database market, but hardly a major threat as a cloud company right now.

They went right for Oracle’s market though with a new on prem system called Outposts, which allows AWS customers to operate on prem and in the cloud using a single AWS control panel or one from VMware if customers prefer. That is the kind of cloud vision that Larry Ellison might have put forth, but Jassy didn’t necessarily see it as going after Oracle or anyone else. “I don’t see Outposts as a shot across the bow of anyone. If you look at what we are doing, it’s very much informed by customers,” he told reporters at a press conference last week.

Yet AWS didn’t reserve its criticism just for Oracle. It also took aim at Microsoft, taking jabs at Microsoft SQL Server, and also announcing Amazon FSx for Windows File Server, a tool specifically designed to move Microsoft files to the AWS cloud.

Google wasn’t spared either when launching Inferentia and Elastic Inference, which put Google on notice that AWS wasn’t going to yield the AI market to Google’s TPU infrastructure. All of these tools and much more were about more than answering customer demand, they were about putting the competition on notice in every aspect of enterprise computing.

Upward growth trajectory

The cloud market is continuing to grow at a dramatic pace, and as market leader, AWS has been able to take advantage of its market dominance to this point. Jassy, echoing Google’s Diane Greene and Oracle’s Larry Ellison, says the industry as a whole is still really early in terms of cloud adoption, which means there is still plenty of marketshare left to capture.

“I think we’re just in the early stages of enterprise and public sector adoption in the US. Outside the US I would say we are 12-36 months behind. So there are a lot of mainstream enterprises that are just now starting to plan their approach to the cloud,” Jassy said.

Patrick Moorhead, founder and principal analyst at Moor Insights & Strategy says that AWS has been using its market position to keep expanding into different areas. “AWS has the scale right now to do many things others cannot, particularly lesser players like Google Cloud Platform and Oracle Cloud. They are trying to make a point with the thousands of new products and features they bring out. This serves as a disincentive longer-term for other players, and I believe will result in a shakeout,” he told TechCrunch.

As for the frenetic pace of innovation, Moorhead believes it can’t go on forever. “To me, the question is, when do we reach a point where 95% of the needs are met, and the innovation rate isn’t required. Every market, literally every market, reaches a point where this happens, so it’s not a matter of if but when,” he said.

Certainly areas like the AWS Ground Station announcement, showed that AWS was willing to expand beyond the conventional confines of enterprise computing and into outer space to help companies process satellite data. This ability to think beyond traditional uses of cloud computing resources shows a level of creativity that suggests there could be other untapped markets for AWS that we haven’t yet imagined.

As AWS moves into more areas of the enterprise computing stack, whether on premises or in the cloud, they are showing their desire to dominate every aspect of the enterprise computing world. Last week they demonstrated that there is no area that they are willing to surrender to anyone.

Floom, the online marketplace and SaaS for independent florists, has raised $2.5 million in a seed funding. The round was round led by Firstminute Capital, and will be used by the London headquartered startup to continue to expand to the U.S., where it already operates in New York and L.A., and to further develop its software offering.

Additional investors include Tom Singh (founder of New Look), Pembroke VCT, Wing Chan (CTO digital experiences of The Hut Group), and Carlos Morgado (former CTO of Just Eat). Morgado has also joined Floom’s board.

Founded by 31-year-old Lana Elie in 2016, Floom bills itself as a curated marketplace for independent florists. Alongside this, the company’s technology platform gives florists the software and tools they need to create and deliver “beautifully crafted bouquets” to customers. It’s this SaaS play that Elie says sets Floom apart from competitors.

“We rely on a network [of florists], like many of the bigger competitors, so that we can offer same-day delivery without the risk of holding stock ourselves,,” she tells me. “But instead of telling the florists what to create and what to hold in stock, we built them an Etsy-like UI to design and deliver beautifully crafted bouquets to our online communities themselves”.

This sees florists provided with a “backend management dashboard” to create, allocate and manage inventory, and to co-ordinate with Floom’s marketplace. The software manages and tracks delivery, too.

“Customers receive more bouquet options, in more areas, by vetted florists, with the ultimate convenience of a seamless check-out and what everyone really wants: confirmation of safe receipt in their loved one’s hand,” explains Elie. “If the final product doesn’t match the picture, they get their money back, something that most competitors can’t offer, but we solved this by relying on the florists to generate the bouquet catalogue themselves”.

On the flower delivery front, Floom’s main competitors are Interflora in the U.K. (owned by 100-year-old conglomerate FTD in the U.S.), as well as 1-800-flowers and Teleflora. “There have been some new players in the flower space, but none solve the problem by creating better technologies,” argues the Floom founder.

“Floom’s not just a flower delivery service but a tech company. I wanted to solve a problem: showing customers all the amazing artisanal florists in their home cities, and making the experience of sending flowers enjoyable and hassle-free. On top of that, we wanted to create a fresh brand that appealed to an audience of my generation… and different from how you might typically think of the flower industry”.

With that said, Elie concedes that there is other florist software in existence, but says it doesn’t really consider the florists as a customer in the same way that Floom does. This is especially true in how the startup understands that the “brand and UI is just as important as functionality”.

“Florists are creative, skilled in a way that I’m definitely not, but when it comes to something like a website build, they’re paying the wrong people much more than they need to build badly UX’d sites,” she adds. “Florists are given no chance to really compete in a world where everything is digital. Building a management tool that speaks to all florists’ consumer facing channels (phone, email, chat, webshop, POS etc) will ultimately mean cost and time savings for the florist, less unnecessary waste for environmental purposes, and better products and delivery experiences for the customer”.

Scooter startup Lime has sought to back peddle on an explanation given by its VP of global expansion late last week when asked why it had hired the controversial PR firm, Definers Public Affairs.

The opposition research firm, which has ties to the Republican Party, has been at the center of a reputation storm for Facebook, after a New York Times report last month suggested the controversial PR firm sought to leverage anti-semitic smear tactics — by sending journalists a document linking anti-Facebook groups to billionaire George Soros (after he had been critical of Facebook).

Last month it also emerged that other tech firms had engaged Definers — Lime being one of them. And speaking during an on stage interview at TechCrunch Disrupt Berlin last Thursday, Lime’s Caen Contee claimed it had not known Definers would use smear tactics.

Yet, as we reported previously, a Definers employee sent us an email pitch in October in which it wrote suggestively that “Bird’s numbers seem off”.

This pitch did not disclose the PR firm was being paid by Lime.

Asked about this last week Contee claimed not to know anything about Definers’ use of smear tactics, saying Lime had engaged the firm to work on its green and carbon free programs — and to try to understand “what were the levers of opportunity for us to really create the messaging and also to do our own research; understanding the life-cycle; all the pieces that are in a very complex business”.

“As soon as we understood they were doing some of these things we parted ways and finished our program with them,” he also said.

However, following the publication of our article reporting on his comments, a Lime spokesperson emailed with what the subject line billed as a “statement for your latest story”, tee-ing this up by writing: “Hoping you can update the piece”.

The statement went on to claim that Contee “misspoke” and “was inaccurate in his description of [Definers] work”.

However it did not specify exactly what Contee had said that was incorrect.

A short while later the same Lime spokesperson sent us another version of the statement with updated wording, now entirely removing the reference to Contee.

You can read both statements below.

As you read them, note how the second version of the statement seeks to obfuscate the exact source of the claimed inaccuracy, using wording that seeks to shift blame in way that a casual reader might interpret as external and outside the company’s control…

Statement 1:

Our VP of Global Expansion misspoke at TechCrunch Disrupt regarding our relationship with Definers and was inaccurate in his description of their work. As previously reported, we engaged them for a three month contract to assist with compiling media coverage reports, limited public relations and fact checking, and we are no longer working with Definers.

What was presented at Disrupt regarding our relationship with Definers and the description of their work was inaccurate. As previously reported, we engaged them for a three month contract to assist with compiling media coverage reports, limited public relations and fact checking, and we are no longer working with Definers.

Despite the Lime spokesperson’s hope for a swift update to our report, they did not respond when we asked for clarification on what exactly Contee had said that was “inaccurate”.

A claim of inaccuracy that does not provide any detail of the substance upon which the claim rests smells a lot like spin to us.

Three days later we’re still waiting to hear the substance of Lime’s claim because it has still not provided us with an explanation of exactly what Contee said that was ‘wrong’.

Perhaps Lime was hoping for a silent edit to the original report to provide some camouflaging fuzz atop a controversy of the company’s own making. i.e. that a PR firm it hired tried to smear a rival.

If so, oopsy.

Of course we’ll update this report if Lime does get in touch to provide an explanation of what it was that Contee “misspoke”. Frankly we’re all ears at this point.

Virtual Reality is in a public relations slump. Two years ago the public’s expectations for virtual reality’s potential was at its peak. Many believed (and still continue to believe) that VR would transform the way we connect, interact, and communicate in our personal and professional lives.

It’s easy to understand why this excitement exists once you put on a head mounted display. While there are still a limited number of compelling experiences, after you test some of the early successes in the field, it’s hard not to extrapolate beyond the current state of affairs to a magnificent future where the utility of virtual reality technology is pervasive.

However, many problems still exist. The all-in cost for state of the art headsets is still out of reach for the mass market. Most ‘high-quality’ virtual reality experiences still require users to be tethered to their desktops. The setup experience for mass market users is lathered in friction. When it comes down to it, the holistic VR experience is a non-starter for most people. We are effectively in what Gartner refers to as the “trough of disillusionment.”

Yet, the virtual reality market has continued its slow march to mass adoption, and there are tangible indicators that suggest we could be nearing an inflection point.

A shift towards sustainable hardware growth

What you do and do not consider a virtual reality display can dramatically impact your view on the state of the VR hardware industry. Head-mounted displays (HMDs) can be categorized in three different ways:

• Screenless viewers — affordable devices that turn smartphones into a VR experience (e.g., Google Glass, Samsung Gear VR, etc.)
• Standalone HMDs — devices that are not connected to a computer and can independently run content (e.g., Oculus Go, Lenovo Mirage Solo, etc.)
• Tethered HMDs — devices that are connected to a desktop computer in order to run content (e.g., HTC Vive, Oculus Pro, etc.)

2018 has seen disappointing progress in aggregate headset growth. The overall market is forecasted to ship 8.9M headsets in 2018, up from an approximate aggregate shipment of ~8.3M in 2017, according to IDC. On the surface, those numbers hardly describe a market at its inflection point.

However, most of the decline in growth rate can be attributed to two factors. First, screenless viewers have seen a significant decline in shipments as device manufacturers have stopped shipping them alongside smartphones. In the second quarter of 2018, 409K screenless viewers were shipped compared to approximately 1M in the second quarter of 2017. Second, tethered VR headsets have also declined as manufacturers have slowed down the pricing discounts that acted as a steroid to sales growth in 2017.

Looking at the market for standalone HMDs, however, reveals a more promising figure. Standalone VR headsets grew 417% due to the global availability of the Oculus Go and Xiaomi Mi VR. Over time, these headsets are going to be the driver of the VR market as they offer significant advantages compared to tethered headsets.

The shift from tethered to standalone VR headsets is significant. It represents a paradigm shift within the immersive ecosystem, where developers have a truly mobile platform that is powerful enough to enable compelling user experiences.

A premium market segment

There are a few names that come to mind when thinking about products that are available for purchase in the VR market: Samsung, Facebook (Oculus), HTC, and Playstation. A plethora of new products from these marquee names —  and products from new companies entering the market —  are opening the category for a new customer segment.

For the past few years, the market effectively had two segments. The first was a “mass market” segment with notorious devices such as the Google Cardboard and the Samsung Gear, which typically sold for under $100 and offered severely constrained experiences to consumers. The second segment was a “pro market” with a few notable devices, such as the HTC Vive, that required absurdly powerful computing rigs to operate, but offered consumers more compelling, immersive experiences.

It’s possible that this new emerging segment will dramatically open up the total addressable VR market. This “premium” market segment offers product alternatives that are somewhat more expensive than the mass market, but are significantly differentiated in the potential experiences that can be offered (and with much less friction than the “pro market”).

The Oculus Go, the Xiaomi Mi VR, and the Lenovo Solo are the most notable products in this segment. They are the fastest growing devices in this segment, and represent a new wave of products that will continue to roll out. This segment could be the tipping point for when we move from the early adopters to the early majority in the VR product adoption curve.

A number of other products have also been released throughout 2018 that fall into this category, such as Lenovo’s Mirage Solo and Xiaomi’s Mi VR. Even more so, Oculus recently announced that  they’ll be shipping a new headset called Quest this spring, which will sell for $399 and will be the most powerful example of a premium device to date. The all-in price range of ~$200–400 places these devices in a segment consumers are already conditioned to pay (think iPad’s, gaming consoles, etc.), and they offer differentiated experiences primarily attributed to the fact that they are standalone devices.

In 1990, Kleiner Perkins rejected 99.4% of the proposals it received, while investing in 12 new companies a year. Those investees made Kleiner Perkins “the most successful financial institution in the history of the world,” boasting “returns of about 40 percent per year, compounded, for coming up on thirty years.”

Nowadays, the Valley’s VC poster child is Y Combinator, who invest in more like 250 companies annually. They’re famously selective, accepting something like 1.5% of applicants, but still noticeably less selective than Kleiner Perkins in its heyday. They invest less money (though not necessarily that much less; KP bought 25% of Netscape for a mere $5 million back in 1994) in more companies.

In 1995, three networks controlled essentially all American television, and made only enough to fill the week; nowadays there is so much TV that you could bingewatch a new scripted series every day of the year. In 1995, the top ten movies of the year were responsible for 14% of the total box office. So far in 2018, the top ten have claimed a full 25% of the total gross. Something similar happened in publishing; the so-called “midlist” was largely replaced by a “bestseller or bust” attitude.

In 1995, if you were a journalist, your readership was dictated almost entirely by who published you. No matter how compelling your piece in the Halifax Daily News may have been, the same number of people would glance at your headline as at the others in that issue, and that number would be drastically smaller than that of any article, no matter how buried, in the New York Times. Now, the relative readership of any article, both between and within publications, is determined mostly by social media sharing, and inevitably follows a power-law curve, such that a surprisingly small number of pieces attract the lion’s share of readers.

What do these fields have in common? The number of “hits” has remained relatively constant, while their value has grown, and the number of “swings” has grown to the point where it is difficult for any person, or even any group, to pay close attention to them all. And the outcomes inevitably follow a power law. So it doesn’t make sense to focus on individual outcomes any more; instead you focus on cohorts, and you think stochastically.

“Stochastic” means “randomly determined,” and your initial inclination may be to recoil — of course producers and investors and publishers aren’t acting randomly! They put enormous amounts of analysis, effort, and intelligence into what they do! Which is true. But I put it to you that as gatekeepers’ power has diminished, and the number of would-be directors, CEOs, and pundits has skyrocketed, while the costs of trying have shrunk — randomness has become a more and more important factor.

It’s easy to cite anecdotes. What if Excite had bought Google when it was offered to them for $1 million? How far were we, really, from a world in which Picplz succeeded and Instagram failed? Any honest success story will include elements of luck, which is, in this context, another word for randomness. My contention is that the world’s larger trends — greater interconnectedness, faster speed, democratized access to technology — make randomness an ever-more-important factor.

This is not automatically a good thing. People talk about “stochastic terrorism,” a.k.a. “The use of mass public communication, usually against a particular individual or group, which incites or inspires acts of terrorism which are statistically probable but happen seemingly at random.” Think of killers who dedicate their attack to ISIS after the fact despite no previous communication with them, or, more generally and contentiously, political violence promoted by broadcasting hatred and extremism.

And it seems that climate change is increasingly a stochastic disaster. Warmer weather means more energy in the atmosphere, which means more volatile behavior, which means more catastrophes like droughts, wildfires, hurricanes. Does climate change cause those things? Not directly. It increases the probability of them happening. It means both more and bigger hits, if you will.

This doesn’t apply to every field of human endeavor. But it seems to apply to essentially every field driven by unusual, extreme successes or failures — to Extremistan, to use Nassim Taleb’s term. Extremistan seems to everywhere be growing more extreme, and there’s no end in sight.

Facebook is still dealing with the fallout from a New York Times report outlining the company’s strategy to fight back against criticism, particularly its work with Definers Public Affairs, an opposition research firm with ties to the Republican Party.

That work included a document that Definers sent to reporters suggesting ties between George Soros and progressive political groups criticizing Facebook. The Times story described the broad strokes of the claims made by Definers, but the document itself has not been shared with the public — until today, when it was published by BuzzFeed.

At this point, the contents aren’t particularly revelatory, but the document is still worth reading, since it’s at the center of the recent controversy.

It’s titled “Freedom From Facebook Potential Funding,” and it begins:

Recently, a number of progressive groups came together to form the Freedom From Facebook campaign which has a six-figure ad budget. It is not clear who is providing the large amount of funding for the campaign but at least four of the groups in the coalition receive funding or are aligned with George Soros who has publicly criticized Facebook. It is very possible that Soros is funding Freedom From Facebook.

The document goes on to point out connections between Soros and several of the groups involved in Freedom From Facebook, and it notes Soros’ public criticism of Facebook and Google. On its own, the document seems “largely innocuous” (as BuzzFeed put it), but it’s become controversial for potentially playing into anti-Semitic conspiracy theories about Soros.

A Freedom From Facebook spokesperson has said that no money from Soros was used to fund the campaign — in fact, Axios reported that its initial funding came from David Magerman, a Pennsylvania-based philanthropist and former hedge fund executive.

According to BuzzFeed, this is one of at least two documents that Definers prepared after Soros made critical remarks about Facebook and Google at Davos.

Meanwhile, CEO Mark Zuckerberg and COO Sheryl Sandberg have denied knowledge of Definers’ work for Facebook, and outgoing head of public policy Elliot Schrage took responsibility for hiring the firm. But Facebook later acknowledged that Sandberg had asked the communications team to research Soros’ financial ties after he criticized the company, and reporting by my colleague Taylor Hatmaker suggests that Sandberg was more aware of Definers’ work than initially acknowledged.

When reached for comment, a Facebook spokesperson pointed us to Schrage’s post and said the company has nothing further to add.