The market for “prompt engineers” to drive AI systems is booming. You could call them wizards casting spells. CC-licensed photo by Louis K.Louis K. on Flickr.
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 10 links for you. Abracadabra: you’re rich! I’m @charlesarthur on Twitter. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
Elon Musk joins call for pause in creation of giant AI ‘digital minds’ • The Guardian
Alex Hern:
»
More than 1,000 artificial intelligence experts, researchers and backers have joined a call for an immediate pause on the creation of “giant” AIs for at least six months, so the capabilities and dangers of systems such as GPT-4 can be properly studied and mitigated.
The demand is made in an open letter signed by major AI players including: Elon Musk, who co-founded OpenAI, the research lab responsible for ChatGPT and GPT-4; Emad Mostaque, who founded London-based Stability AI; and Steve Wozniak, the co-founder of Apple.
Its signatories also include engineers from Amazon, DeepMind, Google, Meta and Microsoft, as well as academics including the cognitive scientist Gary Marcus.
“Recent months have seen AI labs locked in an out-of-control race to develop and deploy ever more powerful digital minds that no one – not even their creators – can understand, predict, or reliably control,” the letter says. “Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable.”
The authors, coordinated by the “longtermist” thinktank the Future of Life Institute, cite OpenAI’s own co-founder Sam Altman in justifying their calls. In a post from February, Altman wrote: “At some point, it may be important to get independent review before starting to train future systems, and for the most advanced efforts to agree to limit the rate of growth of compute used for creating new models.”
The letter continued: “We agree. That point is now.”
If researchers will not voluntarily pause their work on AI models more powerful than GPT-4, the letter’s benchmark for “giant” models, then “governments should step in”, the authors say.
«
It was going so well until Musk signed it, which makes one think maybe it’s not such a problem.
unique link to this extract
iPhone 15 Pro low-energy chip to allow solid-state buttons to work when device is off or out of battery • MacRumors
Tim Hardwick:
»
The iPhone 15 Pro and Pro Max will use a new ultra-low energy microprocessor allowing certain features like the new capacitive solid-state buttons to remain functional even when the handset is powered off or the battery has run out, according to a source that shared details on the MacRumors forums.
The source of this rumour is the same forum member that shared accurate details about the Dynamic Island last year before the iPhone 14 Pro was officially launched, so there is good reason to believe that the following information is reliable.
According to the anonymous source, the new microprocessor will replace Apple’s current super-low energy mode that allows an iPhone to be located via Find My after it has been powered off or for up to 24 hours if its battery has been depleted, and enables Apple Pay Express Mode to be used for up to five hours after the battery has run out.
The new chip will allegedly take over these existing Bluetooth LE/Ultra Wideband functions in addition to powering the solid-state buttons – including an “action” button that replaces the mute switch – when the phone is on, off, or the battery is depleted.
«
An action button replacing the mute switch? I suppose the action could be “mute”. (Or “unmute”.) At this point we might as well do a version of the bad comedian’s joke – “why don’t they make the aircraft out of the same stuff they use for the black box?” – and ask why they don’t power the phone with this low-energy microprocessor.
unique link to this extract
Disney, Microsoft say meh to the metaverse • WSJ
Meghan Bobrowsky:
»
The metaverse, the virtual world that was the hot thing in tech less than two years ago, is facing a harsher reality.
Walt Disney has shut down the division that was developing its metaverse strategies, The Wall Street Journal reported this week. Microsoft recently shut down a social virtual-reality platform it acquired in 2017. And Mark Zuckerberg, who renamed Facebook as Meta Platforms to signal his seriousness about the metaverse, focused more on artificial intelligence on an earnings call last month.
Meanwhile, the price for virtual real estate in some online worlds, where users can hang out as avatars, has cratered. The median sale price for land in Decentraland has declined almost 90% from a year ago, according to WeMeta, a site that tracks land sales in the metaverse.
Meta’s name change in October 2021 spurred excitement about metaverse experiences, products and platforms. But slow user adoption, driven in part by expensive hardware requirements and glitchy tech, and deteriorating economic conditions have put a damper on expectations the metaverse will drive meaningful revenue anytime soon.
«
Fantastic headline. Plus it’s really hard to see Apple seriously unveiling a VR headset at its Worldwide Developers Conference (which is from June 5-9, announced on Wednesday) and expecting anyone to think much of it. The game’s all with AI right now.
unique link to this extract
U.S. Energy Information Administration • EIA – Independent Statistics and Analysis
»
Last year, the U.S. electric power sector produced 4,090 million megawatthours (MWh) of electric power. In 2022, generation from renewable sources—wind, solar, hydro, biomass, and geothermal—surpassed coal-fired generation in the electric power sector for the first time. Renewable generation surpassed nuclear generation for the first time in 2021 and continued to provide more electricity than nuclear generation last year.
Natural gas remained the largest source of U.S. electricity generation, increasing from a 37% share of U.S. generation in 2021 to 39% in 2022. The share of coal-fired generation decreased from 23% in 2021 to 20% in 2022 as a number of coal-fired power plants retired and the remaining plants were used less. The share of nuclear generation decreased from 20% in 2021 to 19% in 2022, following the Palisades nuclear power plant’s retirement in May 2022.
«
Looking at the graph, gas is basically replacing coal, even while energy demand in total is going up, and renewables are filling the demand gap.
unique link to this extract
US court sanctions Google for deleting evidence in antitrust cases • Reuters
Mike Scarcella:
»
Alphabet Inc’s Google LLC intentionally destroyed employee “chat” evidence in antitrust litigation in California and must pay sanctions and face a possible penalty at trial, a US judge ruled on Tuesday.
US District Judge James Donato in San Francisco said in his order that Google “fell strikingly short” in its duties to preserve records. The ruling is part of a multidistrict litigation that includes a consumer class action with as many as 21 million residents; 38 states and the District of Columbia; and companies including Epic Games Inc and Match Group LLC.
The consumers and other plaintiffs are challenging Google’s alleged monopoly for distributing Android mobile applications, allegations that Google has denied. Plaintiffs have claimed aggregate damages of $4.7bn.
The judge asked the plaintiffs’ lawyers by April 21 to provide an amount in legal fees they are seeking as a sanction.
Separately, the plaintiffs will have a chance to urge Donato to tell jurors that Google destroyed information that was unfavorable to it. He said he wants to see “the state of play” at a later stage in the case.
“Google has tried to downplay the problem and displayed a dismissive attitude ill tuned to the gravity of its conduct,” the judge said.
A Google spokesperson on Tuesday said the company has “produced over three million documents, including thousands of chats.”
In a court filing last year, Google’s lawyers said the company took “robust steps to preserve relevant chats.”
«
Apparently Google left it up to participants in chats whether to retain them even after the litigation had started. The judge’s decision is pretty brutal.
unique link to this extract
Twitter is dying • TechCrunch
Natasha Lomas:
»
Twitter is dying.
The value that Twitter’s platform produced, by combining valuable streams of qualification and curiosity, is being beaten and wrung out. What’s left has — for months now — felt like an echo-y shell of its former self. And it’s clear that with every freshly destructive decision — whether it’s unbanning the nazis and letting the toxicity rip, turning verification into a pay-to-play megaphone or literally banning journalists — Musk has applied his vast wealth to destroying as much of the information network’s value as possible in as short a time as possible; each decision triggering another exodus of expertise as more long-time users give up and depart.
Simply put, Musk is flushing Twitter down the sink. I guess now we all know what the dumb meme really meant.
On April Fools Day, the next — perhaps final — stage of the destruction will commence as Musk rips away the last layer of legacy verification, turning up the volume on anyone who’s willing to pay him $7.99 per month to shout over everyone else.
Anyone who was verified under the old (and by no means perfect) system of Twitter verification — which was at least related to who they were (celebrity, expert, journalist, etc.) — will cease to be verified. Assuming they haven’t already deleted their account. Only accounts that pay Musk will display a ‘Blue Check’.
This is just a parody of verification since the blue tick no longer signals any kind of quality. But the visual similarity seems intentional; a dark pattern designed to generate maximum confusion.
«
Initially, Musk complained that the “blue tick” of verified accounts created a feudal system (in fact, it created a reputation system). Then he let people buy into it, diluting the reputational usefulness. Then he’s removing the reputation part. It’s like watching someone set money on fire, though the KLF did it with more style and annoyed fewer people.
unique link to this extract
The deepfake AI porn industry is operating in plain sight • NBC News
Kat Tenbarge:
»
An NBC News review of two of the largest websites that host sexually explicit deepfake videos found that they were easily accessible through Google and that creators on the websites also used the online chat platform Discord to advertise videos for sale and the creation of custom videos.
The deepfakes are created using AI software that can take an existing video and seamlessly replace one person’s face with another’s, even mirroring facial expressions. Some lighthearted deepfake videos of celebrities have gone viral, but the most common use is for sexually explicit videos. According to Sensity, an Amsterdam-based company that detects and monitors AI-developed synthetic media for industries like banking and fintech, 96% of deepfakes are sexually explicit and feature women who didn’t consent to the creation of the content.
Most deepfake videos are of female celebrities, but creators now also offer to make videos of anyone. A creator offered on Discord to make a 5-minute deepfake of a “personal girl,” meaning anyone with fewer than 2 million Instagram followers, for $65.
The nonconsensual deepfake economy has remained largely out of sight, but it recently had a surge of interest after a popular livestreamer admitted this year to having looked at sexually explicit deepfake videos of other livestreamers. Right around that time, Google search traffic spiked for “deepfake porn.”
«
AI and the American Smile • Medium
jenka:
»
“The expectation was, you have to smile eight hours a day,” a woman Baker calls Sofiya tells her. A 41-year-old Russian émigré who had been living in the United States for the past decade, Sofiya “was a proficient English speaker,” Baker writes, but it was in her job as a bank teller that she “came face-to-face with her deficiency in speaking ‘American.’ This other English language, made up of not just words but also facial expressions and habits of conversation subtle enough to feel imagined.
Smiling almost constantly was at the core of her duties as a teller. As she smiled at one customer after another, she would wince inwardly at how silly it felt. There was no reason to smile at her clients, she thought, since there was nothing particularly funny or heartwarming about their interactions. And her face hurt.”
This confrontation with the culture clash of smiling for an Eastern European immigrant in America hits close to home. Which is why seeing the relentless parade of toothy, ahistorical, quintessentially American, “cheese” smiles plastered on the faces of every civilization in the world across time and space was immediately jarring.
It was as if the AI had cast 21st century Americans to put on different costumes and play the various cultures of the world. Which, of course, it had.
«
That sort-of ingratiating smile that the AI of Midjourney puts onto different races and periods really does mark one out.
unique link to this extract
$335,000 pay for ‘AI Whisperer’ jobs appears in red-hot market • BNN Bloomberg
Conrad Quilty-Harper:
»
Everybody is talking about the artificial intelligence behind ChatGPT. Less noticed is a jobs market mushrooming around the technology, where these newly created roles can pay upwards of $335,000 a year.
And for many a computer engineering degree is optional.
They’re called “prompt engineers,” people who spend their day coaxing the AI to produce better results and help companies train their workforce to harness the tools.
Over a dozen artificial intelligence language systems called large language models, or LLMs, have been created by companies like Google parent Alphabet Inc., OpenAI and Meta Platforms Inc.. The technology has moved rapidly from experiments to practical use, with firms like Microsoft Corp. integrating ChatGPT into its Bing search engine and GitHub software development tool.
As the technology proliferates, many companies are finding they need someone to add rigor to their results.
“It’s like an AI whisperer,” says Albert Phelps, a prompt engineer at Mudano, part of consultancy firm Accenture in Leytonstone, England. “You’ll often find prompt engineers come from a history, philosophy, or English language background, because it’s wordplay. You’re trying to distill the essence or meaning of something into a limited number of words.”
Phelps, 29, studied history at the University of Warwick near Birmingham, England, before starting his career as a consultant for banks like Clydesdale Bank and Barclays Plc, helping them solve problems around risk and regulations. A talk from the Alan Turing Institute, a UK-government funded institute for artificial intelligence, inspired him to research AI, leading to his role at Accenture.
«
If the prompts are spells, as Alex Hern once put it, which place will turn out to be Hogwarts?
unique link to this extract
Arm seeks to raise prices ahead of hotly anticipated IPO • Financial Times
Anna Gross, Cheng Ting-Fang and Kana Inagaki:
»
The UK-based group, which designs blueprints for semiconductors found in more than 95% of all smartphones, has recently informed several of its biggest customers of a radical shift to its business model, according to several industry executives and former employees.
These people said Arm planned to stop charging chipmakers royalties for using its designs based on a chip’s value and instead charge device makers based on the value of the device. This should mean the company earns several times more for each design it sells, as the average smartphone is vastly more expensive than a chip.
The changes represent one of the biggest shake-ups to Arm’s business strategy in decades, at a time when SoftBank chief executive Masayoshi Son is seeking to drive up Arm’s profits and attract investors to its impending return to the public markets.
“Arm is going to customers and saying ‘We would like to get paid more money for broadly the same thing’,” said one former senior employee who left the company last year. “What SoftBank is doing at the moment is testing the market value of the monopoly that Arm has.”
«
This is the pricing model that Qualcomm uses, and it’s really not popular with a lot of Qualcomm purchasers, notably Apple, because if you sell expensive smartphones then you’re on the hook for a chunk of the retail price. But if Arm and Qualcomm do it, what option? And does Apple, which has a special licence from Arm to design its own chips (and helped found Arm, back in the day), get an exception?
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
The Overspill: when there's more that I want to say 