Start Up No.1894: ghostwriting tweets for VCs, assisted death by AI, Netflix’s ad plan, the economics of happiness, and more

Not clippy
With AI illustrators all the rage, Microsoft is now offering one that could integrate with Office. Is Clippy back.. as an AI? (Picture by Stable Diffusion: “a paperclip interrupting a presentation”.)

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Sub-hypermobile. I’m @charlesarthur on Twitter. Observations and links welcome.

It’s Friday, so there’s another post due at the Social Warming Substack at about 0845 UK time.

How a ghostwriter makes $200,000 a year writing tweets for top Silicon Valley investors • Business Insider

Mattathias Schwartz spoke to said ghostwriter:


As a ghostwriter, I never log in to a client’s Twitter account. I just write things and they send them out into the world. We use a software tool where I draft tweets, the client sees them, and the client can then choose to send them if and when they want. We often use Trello. I’ll push it into Trello, and their executive assistant will post it.

There isn’t usually much of a filter between what I write and what gets tweeted. Twitter is so fast. You’re either on the zeitgeist or you’re not. Most people will post close to 100% of what I write. If it turns out to be a banger, that’s great. If it’s not a banger, it gets deleted.

I pride myself on not sticking my foot in my mouth. Nothing has turned into a gaffe. There is a set of topics that no matter what you say about them, it leads to people being angry in your replies. And VCs will often choose to engage in those third-rail topics. For example, how many hours should you work? That’s a classic. If a VC feels they’re not getting enough attention, they can just tweet, “You have to work 80 hours a week to be successful.” Everyone will come out to tell you that you’re canceled. It taps into money, privilege, class, ability to sacrifice. People have a lot of emotions about those subjects.

So taking risks can lead to greater attentional rewards, but the precise level of risk I’ll take depends on the client. Some clients don’t care. They’re shock jocks. They’ll tweet anything. Others are more careful. It’s a question of what brand they’re trying to build.

Most VCs who are playing the content game employ a bunch of writers. I know people who make almost seven figures doing this. I’m not there yet. But I definitely make more at this than I do at my day job. I’ve raised tens of millions of dollars for the company I founded, but I’m not allowed to cash out any of my equity. The average founder of a Series C company makes something like $120,000 a year. And you’re expected to cover your travel costs out of pocket. So that work doesn’t actually pay the bills. I can’t afford to stop ghostwriting, because the city where I live is too expensive.


And why does [person] find a market for this? Because SV VCs make their relationships on Twitter now, not in the bars or bistros or barbeques of Sand Hill Road. (OK, there aren’t any there.) It sounds like a lot of money, but that’s from having between 25 and 50 clients. Shifting patterns of work create more abstracted ways to connect. (How soon before an AI is doing this?)
unique link to this extract

AI-generated imagery is the new clip art as Microsoft adds DALL-E to its Office suite • The Verge

James Vincent:


Microsoft is adding AI-generated art to its suite of Office software with a new app named Microsoft Designer.

The app functions the same way as AI text-to-image models like DALL-E and Stable Diffusion, letting users type prompts to “instantly generate a variety of designs with minimal effort.” Microsoft says Designer can be used to create everything from greeting cards and social media posts to illustrations for PowerPoint presentations and logos for businesses.

Essentially, AI-generated imagery looks set to become the new clip art.

The app isn’t ready for a full launch though, and Microsoft is only offering a limited web preview. “We’re inviting people to try it out, give us feedback, and help us make it great,” writes Microsoft vice president Liat Ben-Zur in a blog post. Once it’s ready, Designer will be available as both a free standalone app and a more feature-filled version that will be available to paying Microsoft 365 subscribers.


Super bonus point to Nigel Moss on Twitter, who simply said: “Clipp-E”?
unique link to this extract

How AI could be used to make life and death decisions • MIT Technology Review

Will Douglas Heaven:


A coffin-size pod with Star Trek stylings, the Sarco is the culmination of Philip Nitschke’s 25-year campaign to “demedicalize death” through technology. Sealed inside the machine, a person who has chosen to die must answer three questions: who are you? Where are you? And do you know what will happen when you press that button? 

Here’s what will happen: The Sarco will fill with nitrogen gas. Its occupant will pass out in less than a minute and die by asphyxiation in around five. 

A recording of that short, final interview will then be handed over to the Swiss authorities. Nitschke has not approached the Swiss government for approval, but Switzerland is one of a handful of countries that have legalized assisted suicide. It is permitted as long as people who wish to die perform the final act themselves. 

Nitschke wants to make assisted suicide as unassisted as possible, giving people who have chosen to kill themselves autonomy, and thus dignity, in their final moments. “You really don’t need a doctor to die,” he says. 

Because the Sarco uses nitrogen, a widely available gas, rather than the barbiturates that are typically used in euthanasia clinics, it does not require a physician to administer an injection or sign off on lethal drugs. 

At least that’s the idea. Nitschke has not yet been able to sidestep the medical establishment fully. Switzerland requires that candidates for euthanasia demonstrate mental capacity, Nitschke says, which is typically assessed by a psychiatrist. “There’s still a belief that if a person is asking to die, they’ve got some sort of undiagnosed mental illness,” he says. “That it’s not rational for a person to seek death.”

He believes he has a solution, however. Exit International is working on an algorithm that Nitschke hopes will allow people to perform a kind of psychiatric self-assessment on a computer. In theory, if a person passed this online test, the program would provide a four-digit code to activate the Sarco. “That’s the goal,” says Nitschke. “Having said all that, the project is proving very difficult.” 


You may be able to see that having an algorithm decide whether you’re mentally competent is proving a bit of a stumbling block. Also, this pod idea doesn’t really work for people who don’t have the use of their limbs, which in wasting diseases is often linked with a wish for assisted death. Or maybe you’d have to speak the code, and then it would work? All very sketchy, really, compared to the certainty one feels about having humans involved. This pod thing is too reminiscent of Futurama’s suicide booths.

Plus, nominative determinism prize for having someone surnamed Heaven writing on this topic.
unique link to this extract

Netflix to charge $6.99 a month for ad-supported plan starting Nov. 3 • CNBC

Alex Sherman:


Netflix will charge $6.99 per month for its new advertising-supported tier, which the company will roll out in the U.S. on Nov. 3.

Netflix’s “Basic with ads” tier will include an average of four to five minutes of commercials each hour and won’t give users the ability to download movies and TV series. A limited number of TV series and movies will initially be unavailable due to licensing restrictions.

Ads will be 15 or 30 seconds in length and will play before and during Netflix’s content. Companies will have the ability to prevent ads from appearing on content they deem unsavoury or unsuitable. To help advertisers understand its reach, ratings company Nielsen will use its standard digital audience measurement, Digital Ad Ratings, in the US beginning in 2023.

…Netflix priced the service so that any customer who switches to the ad-supported service from the ad-free basic plan will have a “neutral to positive” effect on the company’s revenue, according to Peters.

That suggests Netflix will get at least $3 a month per user in advertising revenue.

“We want to offer consumers choice and figure out what the best offering is for them,” Peters said during the conference call.

Video resolution for Netflix’s advertising tier will be 720p rather than 1080p, the quality of Netflix’s standard plan that costs $15.49 per month. The company’s basic plan without advertising is $9.99 per month and also has 720p resolution.


Four minutes per hour will seem like nothing to American audiences. US broadcast channels average 12 to 17 (!) minutes per hour. Even in the UK, ad breaks can total up to 8 minutes per hour on “terrestrial” channels.
unique link to this extract

Economists must get more in touch with peoples’ feelings • Tim Harford

Tim Harford:


I once gently teased the happiness research community by suggesting we wouldn’t learn much about how to reform a nation’s economic institutions by asking citizens, “Overall, how rich do you think you are these days, on a scale of 0-10?” The question seems silly and a reminder of how little we really know about wellbeing.

Well, the joke is on me. Perhaps that is precisely the question we should be asking. A recent study by Federica Liberini, Andrew Oswald, Eugenio Proto and Michela Redoano looked at the impact of how people feel about their finances. Liberini and her colleagues looked at a question from a long-running academic survey, Understanding Society: “How well would you say you yourself are managing financially these days?”. Answers varied from 1 (living comfortably) to 5 (finding it very difficult).

The researchers found that people who said they were living comfortably were more likely to support the Remain campaign in the UK. Those who found their finances very difficult tended to sympathise with Vote Leave. Indeed, write the researchers, “UK citizens’ feelings about their incomes were a substantially better predictor of pro-Brexit views than their actual incomes.”

Then there is inequality. Objectively speaking, it is far from clear that income inequality is rising. In the UK, income inequality rose to high levels during the 1980s and has broadly stayed there ever since. Globally, there is no obvious cause for alarm either. Incomes have risen much faster in China and India — two large, poor countries — than in the US or Europe, putting downward pressure on income inequality.

But people’s feelings? They tell a different story. Jon Clifton, the head of Gallup, which has been tracking wellbeing around the world for many years, notes a polarisation in people’s life-evaluations. Compared with 15 years ago (before the financial crisis, smartphones and Covid-19) twice as many people now say they have the best possible life they could imagine (10 out of 10); however, four times as many people now say they are living the worst life they can conceive (0 out of 10). About 7.5% of people are now in psychological heaven, and about the same proportion are in psychological hell.


The Brexit data is a new one on me. But the story of society being pulled apart sounds very familiar.
unique link to this extract

Where is all the book data? • Public Books

Ruby Ray Daily:


BookScan’s influence in the publishing world is clear and far-reaching. To an editor, BookScan numbers offer two crucial data points: (1) the sales history of the potential author, if it exists, and (2) the sales history of comparable, or “comp,” titles. These data points, if deemed unfavorable, can mean a book is dead in the water.

Take it from freelance editor Christina Boys, whom I spoke with over email, and who worked for 20 years as an editor at two of the Big Five publishing houses (Simon & Schuster and Hachette Book Group). Boys told me that BookScan data is “very important” for deciding whether to acquire or pass on a book; BookScan is also used to determine the size of an advance, to dictate the scale of a marketing campaign or book tour, and to help sell subsidiary rights like translation rights or book club rights. “A poor sales history on BookScan often results in an immediate pass,” Boys said.

Clayton Childress, a sociologist at the University of Toronto, came to similar conclusions in his 2012 study of BookScan data, in which he interviewed and observed more than 40 acquisition editors from across the country. Bad book sales numbers can haunt an author “like a bad credit score,” Childress reported, and they can “caus[e] others to be hesitant to do business with them because of past failures.”

According to editors like Boys, the sway of book sales figures has siphoned much of the creativity and originality out of contemporary book publishing. “There’s less opportunity to acquire or promote a book based on things like gut instinct, quality of the writing, uniqueness of an idea, or literary or societal merit,” Boys claimed. “While passion—arguing that a book should be published—still matters, using that as a justification when it’s contrary to BookScan data has become increasingly challenging.”

…Trubek says that BookScan data encourages publishers to keep recycling the same kinds of books that sold well in the past. “I didn’t want to be a publisher who was working that way,” she elaborated.


But you can see the publishers’ point. It’s just like the music business: what you want is surefire hits. From a new act. Failing that, surefire hits from an old act.
unique link to this extract

Eero ends its $30 a year Eero Secure subscription plan • The Verge

Jennifer Pattison Tuohy:


Earlier this month, Eero announced it would be transitioning its paid subscription plans to one flat rate of $9.99 a month or $99.99 a year. Formerly called Eero Secure Plus, Eero Plus is now the only option for users of the mesh Wi-Fi system if they want to access features like parental controls, ad blocking, or any network usage information.

Today, subscribers to the now-defunct cheaper plan, Eero Secure, started to receive notifications that their plan was ending and would be transferred to the new — more expensive — Eero Plus plan starting on November 15th.

A subscription is not required to use Eero devices, but without one, you can’t set parental controls or see how devices behave on your network, including how much data is being used. These are services that most router manufacturers offer for free. For example, Google Nest Wifi offers parental controls and network activity information for free on its mesh network system.


Parental controls are one thing (though you can probably set those on the devices themselves?) but do many people spend any time looking at “how devices behave” on their network, or monitor their data usage? Some crazy plans might limit data, but surely not many – and people on plans like that would be unlikely to be experienced enough to deal with this. (Plus, now they might as well spend the extra money on an unlimited data plan.)

Peculiar move, though. It’s not as if that functionality actually costs Amazon anything at all to implement. It’s inherent in any router: it’s an essential, configurable part of the webserver it’ll have running to show an interface with the user. Money for nothing, really.
unique link to this extract

The amazing power of “machine eyes” • Ground Truths

Eric Topol:


Today’s report on AI of retinal vessel images to help predict the risk of heart attack and stroke, from over 65,000 UK Biobank participants, reinforces a growing body of evidence that deep neural networks can be trained to “interpret” medical images far beyond what was anticipated. Add that finding to last week’s multinational study of deep learning of retinal photos to detect Alzheimer’s disease with good accuracy. In this post I am going to briefly review what has already been gleaned from 2 classic medical images—the retina and the electrocardiogram (ECG)—as representative for the exciting capability of machine vision to “see” well beyond human limits. Obviously, machines aren’t really seeing or interpreting and don’t have eyes in the human sense, but they sure can be trained from hundreds of thousand (or millions) of images to come up with outputs that are extraordinary. I hope when you’ve read this you’ll agree this is a particularly striking advance, which has not yet been actualized in medical practice, but has enormous potential.

First, a review on deep learning from the retina. We should have known a few years back that something was rich (dare I say eye-opening) about the retina that humans, including retinal experts, couldn’t see. While there are far simpler ways to determine gender, it’s a 50-50 toss up for ophthalmologists, which means there are no visible cues to human eyes. But now two models have shown 97% accuracy of gender determination from neural network training. That was just the beginning.


I’ll be extremely picky and say that the Nature abstract doesn’t make clear whether it’s gender or sex that the retinal systems identify (and 97% offers a lot of edges anyway). But what the systems can detect is amazing; even if they’re wrong, it’s worth checking.
unique link to this extract

💊 Mango Markets exploited for $100m • Web3 Pills

Alex Valaitis writes a daily newsletter on the goings-on, which tend to be – oh, you guessed:


Yesterday alone, saw four DeFi hacks accounting for $115m! The vast majority of funds stolen came from a single exploit on a protocol called Mango Markets.

For those that are unfamiliar, Mango is one of the leading decentralized exchanges on the Solana network. In fact, the hack of Mango accounted for close to 10% of the entire DeFi TVL on Solana. 🤯

Once the Mango team realized what was happening, they froze the entire platform to prevent further damage, but at that point it was already too late.

…In many ways, the drama only began with the exploit. In the aftermath, there has already been a wild sequence of events. Most notably, the attackers went into the governance forum for Mango and made a proposal to return a portion of the funds in return for a bounty and the DAO committing to not pursue criminal investigations.

What is most absurd about this entire proposal, was that most of the ‘Yes’ votes in favor of it, were actually placed by the attackers themselves with funds from the exploit!

On the one hand, this situation has become almost comical and feels like an entertaining scene out of a movie. On the other hand, it shows just how immature the Web3 industry still is. Mango Markets was a “top” DeFi protocol built on the 6th largest blockchain by market cap, Solana.

The fact that they were able to suffer this large of an attack reflects poorly on the entire DeFi space. How can participants have confidence when a massive DeFi attack seems to happen every week?

Not to mention, there are a lot of important questions being asked in regards to who was behind this. It’s an open secret in crypto, that some nefarious developers and/or auditors, intentionally leave behind exploitable code. Some have been speculating whether or not this was an inside job of some sort.

While I think it’s dangerous to speculate to that degree right now, it is worth pointing out that this exact vulnerability was laid out in the Mango Markets Discord back in March.


Valaitis then points to someone describing, yup, exactly how the vulnerability could work. It is hilarious that the hacker(s) could, by virtue of the hack they’d carried out, give themselves the authority to pay a bounty and block any further action. Brilliant twist.
unique link to this extract

• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.

Errata, corrigenda and ai no corrida: none notified

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.