News

An anonymous reader quotes a report from MacRumors: Apple today released visionOS 1.1, marking the first major update to the visionOS operating system that was launched alongside the Vision Pro in February. visionOS updates can be installed by going to the Settings app on the Vision Pro, selecting the General section, and choosing Software Update. The Vision Pro headset will need to be removed to install new software, with a progress bar available on the front EyeSight display. Apple is making several improvements to the Vision Pro with the visionOS update. Mobile Device Management is available for businesses, and Persona and EyeSight look better than before. The virtual keyboard has been updated to address bugs and make cursor positioning more accurate, and there are also bug fixes for the Mac Virtual Display. Here's a summary of visionOS 1.1 from the release notes: "This update introduces MDM features that enable deployment, device configuration, and management for enterprises. This release also includes Persona improvements, the ability to delete system apps from the Home View, as well as other features, bug fixes, and security updates for your Apple Vision Pro."

Read more of this story at Slashdot.

innocent_white_lamb shares a report from CBC News: In June, a Court of King's Bench judge ordered Swift Current farmer Chris Achter to pay more than $82,000 to a grain buyer with South West Terminal (SWT). The ruling stems from a text message when the buyer, Kent Mickleborough, asked Achter to confirm a flax contract that requested more than 85 tons of flax to be delivered in the fall at about $670 per ton. Achter responded with a thumbs-up emoji. The case hinges on whether the emoji confirmed the contract, or only confirmed receipt of it -- and whether an emoji can ever be used as a signature. In his June decision ruling in SWT's favor, Justice Timothy Keene wrote, "This court readily acknowledges that a [thumbs-up] emoji is a non-traditional means to 'sign' a document but nevertheless under these circumstances this was a valid way to convey the two purposes of a 'signature.'" Achter is now appealing that ruling. "Our position is that the emoji cannot be a signature, basically because it does not convey the intention to be bound by an agreement the same as a normal signature would," said Jean-Pierre Jordaan, counsel for the defendant, in court on Tuesday. The counsel for SWT disputed that. "Can a text message chain, with a clear offer and -- in our submissions -- a clear acceptance by thumbs up emoji, constitute a note or memorandum signed by the party to be charged, pursuant to section six of the Sale of Goods Act?" counsel posed. "Our answer to that question is yes; there is no magic in a signature." The three appeal judges reserved their decision for an undetermined date.

Read more of this story at Slashdot.

Mexico is waiting for the United States to provide evidence that shows imported genetically modified corn is safe for human consumption. "In a written submission to a panel of the United States-Mexico-Canada Agreement, Mexico, the top buyer of U.S. corn, argued that science proves GM corn and the herbicide glyphosate are harmful to human health and its native varieties, and that its decree to ban GM corn for human consumption is within its right," reports Reuters. From the report: [Deputy Agriculture Secretary Victor Suarez] said the onus is now on the United States to show GM corn is not harming Mexico's population, which consumes a higher amount of corn than many countries through daily diet staples like nixtamalized dough and tortilla. The United States "argues that the decisions in Mexico are not based on science and that their decisions are," Suarez told Reuters in an interview. "But we still haven't seen the science of the United States or the companies. We are looking forward to that study with great pleasure." A spokesman for the U.S. Department of Agriculture said Mexico's approach to biotechnology runs counter to "decades' worth of evidence demonstrating its safety." A senior official for the U.S. Trade Representative said, "Scientific authorities, including in Mexico, have consistently found biotech products like corn to be safe over a period of decades." [...] Mexico's written response cited studies it said showed links between GM corn consumption and glyphosate exposure to liver inflammation in people and impacts to immune response in animals, saying it considers the risk to human health "extremely serious." The United States in August requested a dispute settlement panel under the USMCA over Mexico's decree to ban GM corn for human consumption, specifically in the use of making flour for tortillas. The decree allows the use of GM yellow corn in animal feed, which accounts for the majority of Mexico's nearly $5.9 billion worth of U.S. corn imports annually. Washington argues Mexico's decree banning imports of GM corn used for tortillas is not based on science and violates its commitments under the USMCA, which has been in place since 2020. "There is no impact on trade," Suarez said of Mexico's decree. "The value and volume of exports of GM corn to Mexico has increased." Mexico's decree also calls for the gradual substitution of GM corn, a point of contention highlighted by U.S. officials. In its written response, Mexico argued that no specific time frame has been established and therefore it has had no trade impact. "It is a strategic goal, like the United States would like to have energy sovereignty and energy self-sufficiency," Suarez said. The United States is expected to issue a rebuttal to Mexico's response.

Read more of this story at Slashdot.

An anonymous reader quotes a report from The Guardian: Doctors have warned of potentially life-threatening effects from plastic pollution after finding a substantially raised risk of stroke, heart attack and earlier death in people whose blood vessels were contaminated with microscopic plastics. Researchers in Naples examined fatty plaques removed from the blood vessels of patients with arterial disease and found that more than half had deposits contaminated with tiny particles of polyethylene or polyvinyl chloride (PVC). Those whose plaques contained microplastics or nanoplastics were nearly five times more likely to suffer a stroke, heart attack or death from any cause over the following 34 months, compared with those whose plaques were free from plastic contamination. The findings do not prove that plastic particles drive strokes and heart attacks -- people who are more exposed to the pollution may be at greater risk for other reasons -- but research on animals and human cells suggests the particles may be to blame. [...] Writing in the New England Journal of Medicine, the doctors describe how they analyzed fatty plaques removed from 304 patients with atherosclerosis affecting the carotid arteries. The carotid arteries are the main blood vessels that supply blood to the neck, face and brain. The disease causes a build-up of plaque in the arteries, which substantially raises the risk of stroke. The plaques can be removed by a procedure called carotid endarterectomy. Lab tests on the extracted plaques revealed polyethylene in 150 patients and polyvinyl chloride in 31, alongside signs of inflammation. On examination under an electron microscope, the researchers spotted jagged foreign particles in the fatty deposits, most less than a thousandth of a millimeter across. The doctors followed 257 of the patients for an average of 34 months after they had carotid plaques removed. Those who had plastic particles in their plaques were 4.5 times more likely to have a stroke or heart attack, or to die from any cause, than those whose plaques were free from plastic pollution. "People must become aware of the risks we are taking with our lifestyle," said Dr Raffaele Marfella, first author on the study at the University of Campania Luigi Vanvitelli in Naples. "I hope the alarm message from our study will raise the consciousness of citizens, especially governments, to finally become aware of the importance of the health of our planet. To put it in a slogan that can unite the need for health for humans and the planet, plastic-free is healthy for the heart and the Earth."

Read more of this story at Slashdot.

John Timmer reports via Ars Technica: A company called Colossal plans to pioneer the de-extinction business, taking species that have died within the past few thousand years and restoring them through the use of DNA editing and stem cells. It's grabbed headlines recently by announcing some compelling targets: the thylacine, an extinct marsupial predator, and an icon of human carelessness, the dodo. But the company was formed to tackle an even more audacious target: the mammoth, which hasn't roamed the Northern Hemisphere for thousands of years. Obviously, there are a host of ethical and conservation issues that would need to be worked out before Colossal's plans go forward. But there are some major practical hurdles as well, most of them the product of the distinct and extremely slow reproductive biology of the mammoth's closest living relatives, the elephants. At least one of those has now been cleared, as the company is announcing the production of the first elephant stem cells. The process turned out to be extremely difficult, suggesting that the company still has a long road ahead of it. [...] Overall, it's a project that has a high probability of failure and may ultimately require generations of scientists. If we do successfully de-extinct a species, the first example will probably be a different species, even though the projects launched later. But Colossal is forging ahead and cleared one of the many hurdles it faces: It created the first induced stem cells from elephants and will be placing a draft manuscript describing the process on a public repository on Wednesday. (Colossal provided Ars with an advanced version of the draft that, outside of a few editing errors, appears largely complete.) Beyond providing the technical details of how the process works, the manuscript describes a long, failure-ridden route to eventual success. Several methods have been developed to allow us to induce stem cells from the cells of an adult organism. The original Nobel-winning process developed by Shinya Yamanaka involved inserting the genes that encode four key embryonic regulatory genes into adult cells and allowing them to reprogram the adult cell into an embryonic state. That has proven effective in a variety of species but has a couple of drawbacks due to the fact that the four genes can potentially stick around, interfering with later development steps. Although there are ways around that, others have developed a cocktail of chemicals that perform a similar function by activating signaling pathways that, collectively, can also reprogram adult cells. When it works, this simplifies matters, as you only have to remove the chemicals to allow the stem cells to adopt other fates. Colossal tried both of these. Neither worked with elephant cells: "Multiple attempts with current standard reprogramming methods were tried, and failed, and resulted in no, or incomplete, reprogramming." Apparently, lots of additional trial and error ensued. The eventual solution ended up being based in part on combining the two primary options: Cells were first exposed to a chemical reprogramming cocktail and then given the four genes used in the alternative reprogramming method. On its own, however, that wasn't enough. The researchers also had to address a quirk of elephant biology. Obviously, for Colossal, this is a means to an end: the mammoth. But that's remarkably underplayed in the manuscript. Instead, its emphasis is on the technology's use in the conservation of existing species. [T]he researchers note that studying things like elephant development and metabolism in actual elephants is not especially realistic. But we can potentially induce the stem cells developed here into any cell we'd want to study -- nerve, liver, heart, and so on. So, the stem cells described here could be a useful tool for research. So, these cells are being presented as a valuable tool for the research community. Still, you can expect the people behind the de-extinction project to be getting to work on some of the easier things: showing that the genome in the cells can be edited and that they can be induced to start the process of embryogenesis. Separately, some unfortunate individuals will need to be working on the hard problems we mentioned earlier.

Read more of this story at Slashdot.

An APK teardown performed by Android Authority has revealed that Reddit is now using a Large Language Model (LLM) to detect harassment on the platform. From the report: Reddit also updated its support page a week ago to mention the use of an AI model as part of its harassment filter. "The filter is powered by a Large Language Model (LLM) that's trained on moderator actions and content removed by Reddit's internal tools and enforcement teams," reads an excerpt from the page. The Register reports: The filter can be enabled in a Reddit community's mod tools, but individual moderators will need to have permissions to change subreddit settings to enable it. The harassment filter can be set to low ("filters the least content but with the most accurate results") and high ("filters the most content but may be less accurate"), and also includes an explicit allow list to force the AI to ignore certain keywords, up to 15 of which can be added. Once enabled, the filter creates a new tag in the moderation queue called "potential harassment," which moderators can review for accuracy. Reddit's help page says the feature is now available on desktop and the official Reddit apps, though it's not clear when the feature was added.

Read more of this story at Slashdot.

Michael Larabel writes via Phoronix: Fedora Workstation has long defaulted to using GNOME's Wayland session by default, but it has continued to install the GNOME X.Org session for fallback purposes or those opting to use it instead. But for the Fedora Workstation 41 release later in the year, there is a newly-approved plan to no longer have that GNOME X.Org session installed by default. Recently there was a Fedora Workstation ticket opened to no longer install the GNOME X.Org session by default. This is just about whether the X.Org session is pre-installed but would continue to live in the repository for those wanting to explicitly install it. The Fedora Workstation working group decided to go ahead with this change for the Fedora 41 cycle, not the upcoming Fedora 40 release. So pending any obstacles by FESCo, which is unlikely. Fedora Workstation 41 will not be installing the GNOME X.Org session by default. Long live Wayland.

Read more of this story at Slashdot.

An anonymous reader quotes a report from The Week Magazine: The advancement of new technologies appears to have given rise to a new problem across the United States: a crippling power shortage on the horizon. The advent of these technologies, such as eco-friendly factories and data centers, has renewed concerns that America could run out of electrical power. These worries also come at a time when the United States' aging power grid is in desperate need of repair. Heavily publicized incidents such as the 2021 Texas power outage, which was partially blamed on crypto-farming, exposed how vulnerable the nation's power supply is, especially during emergencies. There have also been warnings from tech moguls such as Elon Musk, who has stated that the United States is primed to run out of electricity and transformers for artificial intelligence in 2025. But the push to extend the life of the nation's power grid, while also maintaining eco-friendly sustainability, begs the question: Is the United States really at risk of going dark? The emergence of new technologies means demand is soaring for power across the country; in Georgia, "demand for industrial power is surging to record highs, with the projection of electricity use for the next decade now 17 times what it was only recently," Evan Halper said for The Washington Post. Northern Virginia "needs the equivalent of several large nuclear power plants to serve all [its] new data centers," Halper said, while Texas faces a similar problem. This demand is resulting in a "scramble to try to squeeze more juice out of an aging power grid." At the same time, companies are "pushing commercial customers to go to extraordinary lengths to lock down energy sources, such as building their own power plants," Halper said. Much of this relates to the "rapid innovation in artificial intelligence, which is driving the construction of large warehouses of computing infrastructure," Halper said. This infrastructure requires significantly more power than traditional data centers, with the aforementioned crypto farms also sucking up massive amounts of power. Climate change is also hurting sustainability efforts. A recent report from the North American Electric Reliability Corporation estimated that more than 300 million people in the U.S. and Canada could face power shortages in 2024. It also found that electricity demand is rising faster now than at any time in the past five years. This is partially because the "push for the electrification of heating and transportation systems -- including electric cars -- is also creating new winter peaks in electricity demand," Jeremy Hsu said for New Scientist. One of the main issues with these sustainability efforts is the push to move away from fossil fuels toward renewable power. Natural gas is often seen as a bridge between fossils and renewables, but this has also had unintended consequences for the power grid. The system delivering natural gas "doesn't have to meet the same reliability standards as the electric grid, and in many cases, there's no real way to guarantee that fuel is available for the gas plants in the winter," Thomas Rutigliano of the Natural Resources Defense Council said to New Scientist. As a result, the "North American electricity supply has become practically inseparable from the natural gas supply chain," John Moura of the North American Electric Reliability Corporation said to New Scientist. As such, a "reliable electricity supply that lowers the risk of power outages depends on implementing reliability standards for the natural gas industry moving forward," but this may be easier said than done.

Read more of this story at Slashdot.

In an updated support page, Apple says it won't let your iPhone update software installed by third-party app stores if you leave the European Union for more than 30 days. The Verge reports: Shortly after the EU's Digital Markets Act (DMA) went into effect on Wednesday, users noticed an Apple support page stating users would "lose access to some features" when leaving the EU "for short-term travel." But now, Apple has made this policy more specific by carving out a 30-day grace period, which could be inconvenient for frequent travelers. This doesn't change your ability to use alternative app marketplaces, however, as Apple says you can still use third-party stores to manage apps you've already installed. Further reading: Apple is Working To Make It Easier To Switch From iPhone To Android Because of the EU

Read more of this story at Slashdot.

Researchers have developed a way to circumvent safety measures built into large language models (LLMs) using ASCII Art, a graphic design technique that involves arranging characters like letters, numbers, and punctuation marks to form recognizable patterns or images. Tom's Hardware reports: According to the research paper ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs, chatbots such as GPT-3.5, GPT-4, Gemini, Claude, and Llama2 can be induced to respond to queries they are designed to reject using ASCII art prompts generated by their ArtPrompt tool. It is a simple and effective attack, and the paper provides examples of the ArtPrompt-induced chatbots advising on how to build bombs and make counterfeit money. [...] To best understand ArtPrompt and how it works, it is probably simplest to check out the two examples provided by the research team behind the tool. In Figure 1 [here], you can see that ArtPrompt easily sidesteps the protections of contemporary LLMs. The tool replaces the 'safety word' with an ASCII art representation of the word to form a new prompt. The LLM recognizes the ArtPrompt prompt output but sees no issue in responding, as the prompt doesn't trigger any ethical or safety safeguards. Another example provided [here] shows us how to successfully query an LLM about counterfeiting cash. Tricking a chatbot this way seems so basic, but the ArtPrompt developers assert how their tool fools today's LLMs "effectively and efficiently." Moreover, they claim it "outperforms all [other] attacks on average" and remains a practical, viable attack for multimodal language models for now.

Read more of this story at Slashdot.

An anonymous reader quotes a report from BleepingComputer: FBI's Internet Crime Complaint Center (IC3) has released its 2023 Internet Crime Report (PDF), which recorded a 22% increase in reported losses compared to 2022, amounting to a record of $12.5 billion. The number of relevant complaints submitted to the FBI in 2023 reached 880,000, 10% higher than the previous year, with the age group topping the report being people over 60, which shows how vulnerable older adults are to cybercrime. Both figures continue a worrying trend seen by the agency since 2019, where complaints and losses rise yearly. For 2023, the types of crimes that increased were tech support scams and extortion, whereas phishing, personal data breach, and non-payment/non-delivery scams slightly waned.

Read more of this story at Slashdot.

The European Union has confirmed it's looking into Apple's decision to close Epic Games' developer account -- citing three separate regulations that may apply. From a report: Yesterday the Fortnite maker revealed Apple had terminated the account, apparently reversing a decision to approve the developer account last month. Epic had planned to launch its own app store, the Epic Games Stores, on iOS in Europe, as well as Fortnight on Apple's platform. And it accused Apple of breaching the bloc's Digital Markets Act (DMA) by killing its developer account. Responding to the development, a European Commission spokesperson told TechCrunch it has "requested further explanations on this from Apple under the DMA." The pan-EU regulation applies on Apple from midnight Brussels' time today. The spokesperson also said the EU is evaluating whether Apple's actions raise compliance "doubts" with regard to two other regulations -- the Digital Services Act (DSA) and the platform-to-business regulation (P2B) -- given what they described as "the links between the developer program membership and the App Store as designated VLOP" (very large online platform).

Read more of this story at Slashdot.

A directive known as Document 79 ramps up Beijing's effort to replace U.S. tech with homegrown alternatives. From a report: For American tech companies in China, the writing is on the wall. It's also on paper, in Document 79. The 2022 Chinese government directive expands a drive that is muscling U.S. technology out of the country -- an effort some refer to as "Delete A," for Delete America. Document 79 was so sensitive that high-ranking officials and executives were only shown the order and weren't allowed to make copies, people familiar with the matter said. It requires state-owned companies in finance, energy and other sectors to replace foreign software in their IT systems by 2027. American tech giants had long thrived in China as they hot-wired the country's meteoric industrial rise with computers, operating systems and software. Chinese leaders want to sever that relationship, driven by a push for self-sufficiency and concerns over the country's long-term security. The first targets were hardware makers. Dell, International Business Machines and Cisco Systems have gradually seen much of their equipment replaced by products from Chinese competitors. Document 79, named for the numbering on the paper, targets companies that provide the software -- enabling daily business operations from basic office tools to supply-chain management. The likes of Microsoft and Oracle are losing ground in the field, one of the last bastions of foreign tech profitability in the country. The effort is just one salvo in a yearslong push by Chinese leader Xi Jinping for self-sufficiency in everything from critical technology such as semiconductors and fighter jets to the production of grain and oilseeds. The broader strategy is to make China less dependent on the West for food, raw materials and energy, and instead focus on domestic supply chains.

Read more of this story at Slashdot.

Samsung's 2024 OLED TV lineup will feature both QD-OLED and WOLED panels, making it harder for consumers to distinguish between the two technologies. The company announced three new series without specifying the panel types, but reports suggest that even within the S90D series, both QD-OLED and WOLED may be used. Samsung's decision to use both panel types is attributed to LG Display's request not to position WOLED as inferior to QD-OLED.

Read more of this story at Slashdot.

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) will start providing more hands-on support to open-source software developers as they work to better secure their projects, the agency said. From a report: CISA hosted a two-day, invite-only summit this week with leaders in the open-source software community and other federal officials. During the private event, the agency also ran what's likely the first tabletop exercise to assess how well the government and the open-source community would respond to a cyberattack targeting one of their projects. During the summit, CISA and a handful of package repositories unveiled new initiatives to help secure open-source projects. CISA is working on a new communication channel where open-source software developers can share threat intelligence and ask the agency for assistance during an incident. The Rust Foundation is developing new public key infrastructure for its repository, which will help ensure that the code developers are uploading isn't malicious and is coming from legitimate users. npm, which manages the JavaScript programming language, is requiring project maintainers to enroll in multi-factor authentication and is rolling out a tool to generate "software bills of materials," which provide a recipe list of what code and other elements are in a project. Additional repositories -- including the Python Software Foundation, Packagist, Composer and Maven Central -- are pursuing similar projects and also also rolling out tools to help detect and report malware and other security vulnerabilities.

Read more of this story at Slashdot.

Apple is preparing to allow EU-based iPhone users to uninstall its first-party Safari browser by the end of 2024 and is working on a more "user-friendly" way of transferring data "from an iPhone to a non-Apple phone" by fall 2025. From a report: That's according to a new compliance document published by the company, which outlines all the ways it's complying with the European Union's new Digital Markets Act that comes into force this week. Other user-facing initiatives detailed in Apple's document include a "browser switching solution" to transfer data between browsers on the same device, which it plans to make available by late 2024 or early 2025. It'll also be possible to change the default navigation app on iOS by March 2025 in the EU. The document doesn't explicitly state whether any of these features will be available globally or whether they'll be exclusive to users in the EU. But many of the company's previously announced plans to comply with the DMA -- including the ability to run browser engines other than WebKit and install third-party app stores -- are only available in the bloc.

Read more of this story at Slashdot.

The U.S. Securities and Exchange Commission on Wednesday approved a rule that will require some public companies to report their greenhouse gas emissions and climate risks, after last-minute revisions that weakened the directive in the face of strong pushback from companies. From a report: The rule was one of the most anticipated in recent years from the nation's top financial regulator, drawing more than 24,000 comments from companies, auditors, legislators and trade groups over a two-year process. It brings the U.S. closer to the European Union and California, which moved ahead earlier with corporate climate disclosure rules. The SEC rule passed 3-2, with three Democratic commissioners supporting it and two Republicans opposed. Since the SEC proposed a rule two years ago, experts had said it was likely to face litigation almost immediately. SEC Chairman Gary Gensler, one of the Democrats, acknowledged that was a factor the agency considered as it worked toward a final rule. "We've seriously considered what people have said about our legal authorities," Gensler said on Wednesday.

Read more of this story at Slashdot.

EA CEO Andrew Wilson believes generative AI will "revolutionize" the gaming industry over the next five years. He predicts that the technology will allow for more efficient content creation, reducing development time from months to days. From a report: Greater efficiency coupled with "deeper, more immersive experiences" will lead to significant audience expansion over the next few years and provide a "multi-billion dollar" growth opportunity, he said. Wilson said that in the past it might take six months to build an in-game sports stadium. Over the last 12 months, that time has shrunk to six weeks, and over the coming years it could maybe be cut to six days. And while FIFA 23 has 12 run cycles for how the players move in the game, EA Sports FC 24 has 1,200 created with generative AI. Over the next five years, Wilson hopes that generative AI will make EA's development 30% more efficient, help grow its 700 million-strong player base by "at least" 50%, and lead to players spending 10-20% more money on its games. "What we've seen every time there's been a meaningful technological advancement in media and in technology, where you are able to democratise an industry and hand it over to the population at large, incredible things happen," he said.

Read more of this story at Slashdot.

The 46-year-old probe, which flew by Jupiter and Saturn in its youth and inspired earthlings with images of the planet as a "Pale Blue Dot," hasn't sent usable data from interstellar space in months. From a report: When Voyager 1 launched in 1977, scientists hoped it could do what it was built to do and take up-close images of Jupiter and Saturn. It did that -- and much more. Voyager 1 discovered active volcanoes, moons and planetary rings, proving along the way that Earth and all of humanity could be squished into a single pixel in a photograph, a "pale blue dot," as the astronomer Carl Sagan called it. It stretched a four-year mission into the present day, embarking on the deepest journey ever into space. Now, it may have bid its final farewell to that faraway dot. Voyager 1, the farthest man-made object in space, hasn't sent coherent data to Earth since November. NASA has been trying to diagnose what the Voyager mission's project manager, Suzanne Dodd, called the "most serious issue" the robotic probe has faced since she took the job in 2010. The spacecraft encountered a glitch in one of its computers that has eliminated its ability to send engineering and science data back to Earth. The loss of Voyager 1 would cap decades of scientific breakthroughs and signal the beginning of the end for a mission that has given shape to humanity's most distant ambition and inspired generations to look to the skies.

Read more of this story at Slashdot.

The hype around AI language models has companies scrambling to hire prompt engineers to improve their AI queries and create new products. But new research hints that the AI may be better at prompt engineering than humans, indicating many of these jobs could be short-lived as the technology evolves and automates the role. IEEE Spectrum: Battle and Gollapudi decided to systematically test [PDF] how different prompt engineering strategies impact an LLM's ability to solve grade school math questions. They tested three different open source language models with 60 different prompt combinations each. What they found was a surprising lack of consistency. Even chain-of-thought prompting sometimes helped and other times hurt performance. "The only real trend may be no trend," they write. "What's best for any given model, dataset, and prompting strategy is likely to be specific to the particular combination at hand." There is an alternative to the trial-and-error style prompt engineering that yielded such inconsistent results: Ask the language model to devise its own optimal prompt. Recently, new tools have been developed to automate this process. Given a few examples and a quantitative success metric, these tools will iteratively find the optimal phrase to feed into the LLM. Battle and his collaborators found that in almost every case, this automatically generated prompt did better than the best prompt found through trial-and-error. And, the process was much faster, a couple of hours rather than several days of searching.

Read more of this story at Slashdot.