Twenty years on, a version of the Star Wars films that aired in Chile is delighting the internet. CC-licensed photo by Gustavo Rivas Valderrama 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.
There’s another post coming this week at the Social Warming Substack on Friday at 0845 UK time. Free signup.
A selection of 10 links for you. Refreshing. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. Observations and links welcome.
Google’s ‘March 2024 core update’ fights back against site spammers • Fast Company
Harry McCracken:
»
Known as the March 2024 Core Update, this round of fixes builds on algorithmic tweaks the company began implementing in 2022 to prevent questionable sites from competing with the useful ones people turn to a search engine to find. In total, Google says, these adjustments should reduce the amount of “low-quality, unoriginal content” by 40%.
Google already penalized sites that used AI to churn out vast amounts of content that was willfully lousy but highly optimized to rank well in its results. With the advent of large language models such as OpenAI’s GPT-4 and Google’s own Gemini, it’s never been easier to stuff a site with AI-generated material. But rather than target sites specifically for harnessing AI in such efforts, Google now says it will focus on curtailing low-grade, high-SEO content regardless of the techniques involved.
“I think generative AI is actually a really valuable tool for creators, and there’s nothing wrong with using it to create the content you create for your users,” says Pandu Nayak, a Google Search VP overseeing quality and ranking. “The problem is when you start creating content at scale not with the idea of serving your users, but with the idea of targeting search ranking.” (Whether the revised policy mentions automation or not may be a wash: It’s tough to imagine anyone who’s mass-producing web pages without regard to their quality not relying on AI to do most of the work.)
«
Hard to think this is an arms race that ends well for Google.
unique link to this extract
Star Wars meets beer ads: George Lucas’ legal battle with Chilean broadcaster • BNN
Geeta Pillai:
»
During the December 2003 broadcasts of several Star Wars films, viewers were treated to an unconventional advertising strategy by channel 13 in collaboration with Cerveza Cristal, one of Chile’s most popular beer brands. Instead of traditional ad breaks, the channel inserted 30-second commercials directly into the movies. These ads were crafted to appear as continuations of the scenes they interrupted, integrating them so smoothly that they seemed to be part of the original films. One ad featured Obi-Wan Kenobi opening a chest to reveal a stash of Cerveza Cristal, complete with rock music and the brand’s logo, while another showed Emperor Palpatine pulling out a beer bottle instead of a lightsaber.
The discovery of these edits has elicited laughter and surprise among the Star Wars community, with some fans expressing a newfound interest in watching these uniquely altered versions of the films. This incident, however, was not taken lightly by Star Wars creator George Lucas. In 2004, Lucas filed a grievance with the Chilean Council for Self-Regulation and Advertising Ethics, leading to a judgment in Lucas Films’ favor. The council decreed that the commercials were not to be aired again. Despite this ruling, the channel and Cerveza Cristal partnered once more in 2004, embedding the beer into scenes from other popular movies like American Beauty, Notting Hill, and Gladiator.
«
These are absolutely the funniest things you will see all day. People have been collecting them all day, mostly via the Twitter user Windy. They’re collected in his thread, or this article.
And they are the best laugh you’ll have all day; possibly all week. Or longer. And it’s such an improvement on the originals.
unique link to this extract
Satellite to ‘name and shame’ worst oil and gas methane polluters • The Guardian
Damian Carrington:
»
A washing-machine-sized satellite is to “name and shame” the worst methane polluters in the oil and gas industry.
MethaneSat is scheduled to launch from California onboard a SpaceX rocket on Monday at 2pm local time (22:00 GMT). It will provide the first near-comprehensive global view of leaks of the potent greenhouse gas from the oil and gas sector, and all of the data will be made public. It will provide high-resolution data over wider areas than existing satellites.
Methane, also called natural gas, is responsible for 30% of the global heating driving the climate crisis. Leaks from the fossil fuel industry are a major source of human-caused emissions and stemming these is the fastest single way to curb temperature rises.
MethaneSat was developed by the Environmental Defense Fund (EDF), a US NGO, in partnership with the New Zealand Space Agency and cost $88m to build and launch. Earlier EDF measurements from planes show methane emissions were 60% higher than calculated estimates published by US authorities and elsewhere.
More than 150 countries have signed a global methane pledge to cut their emissions of the gas by 30% from 2020 levels by 2030. Some oil and gas companies have made similar pledges, and new regulations to limit methane leaks are being worked on in the US, EU, Japan and South Korea.
The EDF’s senior vice-president, Mark Brownstein, said: “MethaneSat is a tool for accountability. I’m sure many people think this could be used to name and shame companies who are poor emissions performers, and that’s true. But [it] can [also] help document progress that leading companies are making in reducing their emissions.”
«
Name and shame, name and praise. It all helps.
We tested Amazon’s new AI shopping chatbot. It’s not good • The Washington Post
Shira Ovide:
»
Amazon’s chatbot [called Rufus] doesn’t deliver on the promise of finding the best product for your needs or getting you started on a new hobby.
In one of my tests, I asked what I needed to start composting at home. Depending on how I phrased the question, the Amazon bot several times offered basic suggestions that I could find in a how-to article and didn’t recommend specific products.
Another time, the Amazon bot suggested items such as a small compost bin, compost bin liners, a garden fork and a compost thermometer.
Compost fans may notice that the first two suggestions were appropriate for collecting compost scraps in your kitchen. The latter two were for making a backyard compost pile. Amazon’s bot appeared to conflate two different needs.
When I clicked the suggestions the bot offered for a kitchen compost bin, I was dumped into a zillion options for countertop compost products. Not helpful.
Because the Amazon chatbot typically shows you a handful of choices, it might feel better than not knowing what product you want and being deluged with a flood of options on Amazon.
Still, when the Amazon bot responded to my questions, I usually couldn’t tell why the suggested products were considered the right ones for me. Or, I didn’t feel I could trust the chatbot’s recommendations.
I asked a few similar questions about the best cycling gloves to keep my hands warm in winter. In one search, a pair that the bot recommended were short-fingered cycling gloves intended for warm weather.
«
AI invents quote from real person in article by Bihar news site: a wake-up call? • The Quint
Karan Mahadik:
»
At first glance, an article about Meta’s AI chatbot that was published on Patna-based news portal Biharprabha reads like a regular 600-word news report that delves into the history of the AI bot, the controversy surrounding its responses, and the concerns raised, in particular, by Dr Emily Bender, a “leading AI ethics researcher”.
“The release of BlenderBot 3 demonstrates that Meta continues to struggle with addressing biases and misinformation within its AI models,” Dr Emily Bender is quoted as saying in the article titled ‘Meta’s AI Bot Goes Rogue, Spews Offensive Content’ published on 21 February.
But it turns out that the real Dr Emily Bender never actually said it. The entire quote was fabricated and misattributed to her in the article that was generated using an AI tool, specifically Google’s Large Language Model (LLM) known as Gemini.
Confirming this with The Quint, Dr Bender said that she “had no record of talking to any journalist from Biharprabha.”
«
Bender is the professor (of linguistics, at the University of Washington) who came up with the phrase “stochastic parrot” to describe LLMs: “stochastic” because it’s probability-based, “parrot” because it says the things without knowing their meaning. Ironically, the quote wrongly attributed to her is the sort of thing she probably would have said.
unique link to this extract
iOS 17.4 won’t remove Home Screen web apps in the EU after all • 9to5Mac
Chance Miller:
»
Apple’s decision to remove Home Screen web apps, also known as progressive web apps or PWAs, faced a lot of criticism. The Open Web Advocacy organization, for example, said “entire categories of apps will no longer be viable on the web as a result” of the change. There were also reports the EU was going to investigate the decision.
At the time, Apple explained that it would have to build an “entirely new integration architecture that does not currently exist in iOS” to address the “complex security and privacy concerns associated with web apps using alternative browser engines.” This, the company said, “was not practical to undertake given the other demands of the DMA and the very low user adoption of Home Screen web apps.”
With [this] announcement, Apple has reversed course and said that Home Screen web apps will continue to exist as they did pre-iOS 17.4 in the European Union. “This support means Home Screen web apps continue to be built directly on WebKit and its security architecture, and align with the security and privacy model for native apps on iOS,” Apple explains today.
This means that all Home Screen web apps will still be powered by WebKit, regardless of whether the web app is added using Safari or not – exactly as it works today and has for years.
«
So it was “not practical” and then it became practical? Hmm.
unique link to this extract
Behind Formula 1’s velvet curtain • Road and Track (archived version)
Kate Wagner:
»
The sound came before the machine and then the machine blurred by and disappeared over the elegant hill, singing. By the second sprint shootout, even though I’ve watched F1 for a few years now, I had no idea what was going on without 10 split-screen views and a guy yelling in my ear. The cards fell where they fell: Max in first as usual, followed by Leclerc, but then, unexpectedly, Alex Albon.
After the second sprint, the INEOS folks informed the journalists that we needed to leave early in order to avoid traffic and make it to dinner on time, where, apparently there would be a special guest. Frustrated, I returned to watching the cars as they started up again, knowing that the drivers were pushing them to their limits, engrossed in their personal kaleidoscope of motion and color.
[Lewis] Hamilton was in one of them. In the last shootout, he drove differently than before. A great verve frayed the lines he was making, something we can only call effort, push. Watching him, I understood what was so interesting about this sport, even though I was watching it in its most bare-bones form—cars going around in circles. The driver is the apotheosis of quick-moving prowess, total focus and control. The car is both the most studied piece of human engineering, tuned and devised in lab-like environments and at the same time a variable entity, something that must be wrestled with and pushed. The numbers are crunched, the forms wind-tunneled. And yet some spirit escapes their control, and that spirit is known only by the driver.
«
Wagner is a cycling journalist, but was given a freebie to go and watch the F1 race in Texas with the sponsors, INEOS. She was stunned by the indifferent affluence on show, and said so. Which is why the article was quickly taken down by Road & Track. Which is a pity: this extract shows the insight that she brings to bear on what seems unbearably dull when seen on TV.
Might be a while before she gets sent on another freebie to F1. But anyone looking for a good writer will remember her name. (Thanks Mark C for the link.)
unique link to this extract
Wisdom of the Silicon Crowd: LLM ensemble prediction capabilities match human crowd accuracy • ArXiv
Philip Tetlock et al:
»
Human forecasting accuracy in practice relies on the ‘wisdom of the crowd’ effect, in which predictions about future events are significantly improved by aggregating across a crowd of individual forecasters. Past work on the forecasting ability of large language models (LLMs) suggests that frontier LLMs, as individual forecasters, underperform compared to the gold standard of a human crowd forecasting tournament aggregate.
In Study 1, we expand this research by using an LLM ensemble approach consisting of a crowd of twelve LLMs. We compare the aggregated LLM predictions on 31 binary questions to that of a crowd of 925 human forecasters from a three-month forecasting tournament. Our main analysis shows that the LLM crowd outperforms a simple no-information benchmark and is statistically equivalent to the human crowd. We also observe an acquiescence effect, with mean model predictions being significantly above 50%, despite an almost even split of positive and negative resolutions.
Moreover, in Study 2, we test whether LLM predictions (of GPT-4 and Claude 2) can be improved by drawing on human cognitive output. We find that both models’ forecasting accuracy benefits from exposure to the median human prediction as information, improving accuracy by between 17% and 28%: though this leads to less accurate predictions than simply averaging human and machine forecasts.
Our results suggest that LLMs can achieve forecasting accuracy rivaling that of human crowd forecasting tournaments: via the simple, practically applicable method of forecast aggregation. This replicates the ‘wisdom of the crowd’ effect for LLMs, and opens up their use for a variety applications throughout society.
«
Tetlock, if the name isn’t familiar, is the man who coined “superforecasting”; this is a look at how good LLMs might be at the task. (Not better than the best humans, seems to be the answer.) You can read the paper. The full list of questions that were posed to the humans and the LLMs is on page 20; they were asked in late 2023, and many have deadlines expiring in January 2024.
unique link to this extract
Delusional self-belief is a superpower… until it’s a disaster • The Ruffian
Ian Leslie:
»
Great things are often accomplished by irrational people for irrational reasons. The rational move for Churchill in May 1940 was to pursue a peace agreement with Germany. In 1997, no Prime Minister without Tony Blair’s luxuriously proportioned ego would have believed it possible to lead a successful peace process in Northern Ireland. Failures of political leadership often stem from leaders without the necessary grandeur of self-conception to really lead – from recent British history, Gordon Brown, Teresa May and Rishi Sunak spring to mind. Leaders without this special sauce tend to flounder around without direction; leaders with too much of it become Liz Truss.
If you’re trying to spot future political stars, look for individuals who display some delusions of grandeur but who aren’t in thrall to them. Similarly, when trying to predict how a political leader will behave, you should factor in the likelihood they are more optimistic about their prospects and abilities than any sane person would be. I often see commentators assuming that a leader’s assessment of the landscape is similar to their’s. This is usually a mistake, and it’s the one I made when I assumed that Joe Biden was unlikely to run for a second term.
Of course, he may still step down, but the fact that we’ve got this close to an election without him doing so is not what I would have predicted when I wrote about his inauguration speech. I assumed that having slain the dragon, he would retire, nobly, to Delaware. In fact, it wasn’t until late in 2022 that his intention to run again became unmistakably clear to me. As soon as it did, I realised I’d made the elementary error of assuming that top-level politicians see the world in the way the rest of us do. To me and other observers, it seemed obvious that he would be too old to run and win in 2024, and be a competent second-term president. Surely Biden would see that too?
No. Joe Biden ran two failed presidential campaigns and didn’t even come close to winning – and still believed he should take another shot, even when nobody else did. He wanted to run in 2016 and was eased out of the way by the Obamas, who thought Clinton was a better bet. Throughout it all, Biden kept believing he could and should be president, and eventually the world came around to where he had been in his mind for fifty years.
So if you’re Joe Biden, of course you believe that you can and should win a second term. Indeed, you believe that you’re the only person in America capable of defeating Trump and governing a divided nation.
«
This is a terrific essay which makes you realise why these politicians do things that to anyone else smack of idiocy. And hidden in there is a fascinating little what-if: what if Biden had been the Democrat candidate in 2016? Would he have repeated VP Gore’s 2000 failure? Or won, and followed on in 2020, giving us a new candidate this year?
unique link to this extract
‘The internet is an alien life form’: how David Bowie created a market for digital music • The Guardian
Eamonn Forde:
»
In early 1998, Virgin Records/EMI had made Massive Attack’s Mezzanine available for streaming in full online at the same time as its physical release, albeit previewing it track-by-track over several weeks. At the time, the British Phonographic Industry (BPI) cautioned against this, suggesting that streaming experiments could increase the possibility of albums being pirated by tech-savvy individuals and burned to CD. This did not stop other major labels or their acts from occasionally experimenting. Both Def Leppard and Red Hot Chili Peppers made their latest albums, respectively Euphoria and Californication, available to stream in full on 4 June 1999, four days before the records would be in the shops. “Getting airplay is getting airplay, you just have to define air,” said Bob Merlis of Warner Bros, the Chili Peppers’ label. “We felt good about this since it was not downloadable.”
But the Bowie album release [Hours – “far from his best album, and not even his best album of the 1990s”] was designed to be a significant step forward. In 1999, he was interviewed by Jeremy Paxman for BBC Newsnight and talked about his career, his art and, most invigoratingly for him, the internet. The 16-minute interview is still available on the BBC website and is frequently shared, especially since Bowie’s death in January 2016, as evidence of his startling prescience with regard to the impact the internet would have on art, politics and society.
“I don’t think we’ve even seen the tip of the iceberg,” he told a wearily cynical Paxman. “I think the potential of what the internet is going to do to society, both good and bad, is unimaginable. I think we’re actually on the cusp of something exhilarating and terrifying.” Paxman, in his arch way, suggested it was just “a tool”, which saw Bowie spring into action. “No, it’s not,” he said. “No – it’s an alien life form!”
He went on to say that the internet would completely change the dynamics of consumption: “The interplay between the user and the provider will be so in simpatico it’s going to crush our ideas of what mediums are all about.”
«
And of course, he was proved right. What we lost when we lost Bowie: vision.
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 
”(PWAs) So it was “not practical” and then it became practical? Hmm”
They simply brought back the old PWA support which uses Safari/WebKit. Probably just a configuration/build flag. Until someone complains and Vestager threatens with a 50 billion euro fine, of course 🙂
EU heard of PWAs for the first time now and they are also not mentioned in the DMA, so there’s probably very little they can do. Effectively nobody uses PWAs.
The truth is that nobody outside Apple’s iOS engineering team has any idea whatsoever of the engineering effort and security/privacy implications of this. It probably isn’t something you can just throw together in a few days – for roughly one billion users.
I do not like the idea that EU commission, guided by Spotify and a handful companies, start to dictate how iPhone should or should not work.
It’s not Apple’s fault in any way that France and Germany are in the digital Stone Age – DMA is all about “fixing” that. Ain’t going to happen.
Ah, not one billion of course since it would (for now) be for EU users only.
But added complexity to the system anyway, for a huge number of users. Complexity always means new attack vectors, for very little user benefit in this case.