SoylentNews

upstart writes:

Even developers who want to pay for the API are having trouble:

Twitter has finally shut off its free API and, predictably, it's breaking a lot of apps and websites. The company had previously said it would cut off access in early February, but later delayed the move without providing an updated timeline.

But, after announcing its new paid API tiers last week, the company seems to have started cutting off the thousands of developers relying on its free developer tools. Over the last couple days, a number of app makers and other services have reported that the Twitter API is no longer functioning. Mashable reported the shutoff seems to have started Tuesday morning, though many developers are still trying to understand what's happening as Twitter doesn't seem to have communicated with most developers about the changes.

The ending of Twitter's free API comes after the company abruptly changed its rules to ban third-party Twitter clients as part of a larger shakeup of its developer strategy. But, as we've previously reported, third-party clients were only a small fraction of the developers, researchers, bot makers and others who relied on Twitter's APIs.

[...] All of these issues are further complicated by the fact that Twitter seems to have communicated very little with any of its developers about these changes or what they mean. Most of the employees who worked in developer relations were cut during the company's mass layoffs. And the company's developer forums are filled with posts from confused developers looking for answers. The company no longer has a communications team, and its press email auto-responds with a poop emoji.

upstart writes:

The first large-scale comparison of mutation rates gives insights into how quickly species can evolve:

Now, a massive analysis of 68 diverse vertebrate species, from lizards and penguins to humans and whales, has made the first large-scale comparison of the rates at which species mutate — a first step toward understanding how quickly they can evolve. The findings, published in the journal Nature, unearthed surprising insights into how the tempo for mutations can change and what sets that pace.

The paper roughly "doubles the amount of mutation-rate estimates we have," said Michael Lynch, an evolutionary biologist at Arizona State University who was not involved in the study. Now we have a "better idea of the amount of variation within vertebrates."

With this extensive data, biologists can begin to answer questions about which traits most influence mutation rates and the pace of evolution. "There are things that affect the rate of evolution, [but] we don't know all of them," said Patricia Foster, a professor emerita of biology at Indiana University who was not involved in the study. "This is the start."

[...] If they found a mutation in around 50% of an offspring's DNA, they concluded that it was likely a germline mutation — one inherited through either the mother's egg or the father's sperm. Natural selection can act directly on such a mutation. Less frequent mutations were deemed to have happened spontaneously in tissues outside the germline; they were less relevant to evolution because they wouldn't get passed on.

(Surprisingly often, mismatches in the family trios told the researchers that the fathers listed by the zoos were unrelated to the babies. Zoo representatives would often shrug at this news and say there might have been two males in the cage. "Yeah, well, the other one is the winner," Bergeron would joke.)

In the end, the researchers had 151 usable trios, representing species as physically, metabolically and behaviorally diverse as massive killer whales, tiny Siamese fighting fish, Texas banded geckos and humans. They then compared the species' mutation rates with what we know about the behaviors and characteristics called their life history. They also considered a statistical measure for each species called the effective population size, which roughly corresponds to how many individuals are needed to represent the genetic diversity. (For example, although the human population today is 8 billion, scientists usually estimate our effective population size to be around 10,000 or fewer.) Bergeron and her colleagues looked for patterns of associations in the numbers.

Freeman writes:

https://arstechnica.com/information-technology/2023/04/meta-introduces-ai-model-that-can-isolate-and-mask-objects-within-images/

On Wednesday, Meta announced an AI model called the Segment Anything Model (SAM) that can identify individual objects in images and videos, even those not encountered during training, reports Reuters.

According to a blog post from Meta, SAM is an image segmentation model that can respond to text prompts or user clicks to isolate specific objects within an image. Image segmentation is a process in computer vision that involves dividing an image into multiple segments or regions, each representing a specific object or area of interest.

The purpose of image segmentation is to make an image easier to analyze or process. Meta also sees the technology as being useful for understanding webpage content, augmented reality applications, image editing, and aiding scientific study by automatically localizing animals or objects to track on video.

Related:
MIT's Computer Vision (CV) Algorithm Identifies Images Down to the Pixel (20220424)
NVIDIA Research's GauGAN AI Art Demo Responds to Words (20211130)
Ask Soylent: Beginning in Artificial Intelligence Methods (20150629)

Original Submission

upstart writes:

Palantir's Plan to Decipher the Mysteries of Long Covid:

65 million people are still suffering from long Covid, the mysterious cocktail of symptoms that persist in some patients more than 12 weeks after an initial infection. Researchers are still working to understand this illness, but it's been slow progress so far.

This is because long Covid is not just a medical problem—it's also a data problem, says Indra Joshi, director of health, research, and artificial intelligence at Palantir, which specializes in analyzing big data.

Before the pandemic, US hospitals kept their data to themselves, Joshi told WIRED Health this March, making it difficult for policymakers and researchers to identify patterns of disease occurring across the country. That's why Palantir worked with the US medical research agency, the National Institutes of Health, to create what Joshi describes as one of the largest collections of Covid-19 health records in the world.

The National COVID Cohort Collaborative, aka N3C, is essentially a giant, collaborative database, enabling clinicians and researchers to study the deidentified data of people suffering from Covid-19 or related conditions. "If you're diagnosed with Covid now, your data goes into this enclave," says Joshi, explaining that N3C now includes 2.1 billion clinical observations. The data enclave also encourages clinicians to enter data in a standardized format, making it easy for their insights to be compared with data collected from other US hospitals.

By harmonizing all this data, the N3C acts as a collective pool of information that researchers can dip into to try to find consensus on the ongoing mysteries of long Covid: What exactly are the symptoms? What treatments are people receiving? And how are they responding to those treatments? Already the N3C data has helped better define the symptoms that make up long Covid. It has also revealed that Black and Hispanic Americans, in comparison to white patients, appear to experience more symptoms and health problems related to long Covid, but are less likely to be diagnosed.

Original Submission

owl writes:

https://www.righto.com/2020/06/die-shrink-how-intel-scaled-down-8086.html

The revolutionary Intel 8086 microprocessor was introduced 42 years ago this month so I've been studying its die.1 I came across two 8086 dies with different sizes, which reveal details of how a die shrink works. The concept of a die shrink is that as technology improved, a manufacturer could shrink the silicon die, reducing costs and improving performance. But there's more to it than simply scaling down the whole die. Although the internal circuitry can be directly scaled down,2 external-facing features can't shrink as easily. For instance, the bonding pads need a minimum size so wires can be attached, and the power-distribution traces must be large enough for the current. The result is that Intel scaled the interior of the 8086 without change, but the circuitry and pads around the edge of the chip were redesigned.

[...] The photo above shows the two 8086 dies at the same scale. The two chips have identical layout in the interior,7 although they may look different at first. The chip on the right has many dark lines in the middle that don't appear on the left, but this is an artifact. These lines are the polysilicon layer, underneath the metal; the die on the left has the same wiring, but it is very faint. I think the newer chip has a thinner metal layer, making the polysilicon more visible.

The magnified photo below shows the same circuitry on the two dies. There is an exact correspondence between components in the two images, showing the circuitry was reduced in size, not redesigned. (These photos show the metal layer on top of the chip; some polysilicon is visible in the right photo.)

upstart writes:

NASA's TEMPO Mission Could Bring You Hyper-Local Pollution Alerts:

An advanced tool for monitoring air pollution from space is set to launch on Friday morning, and it'll be hosted on a commercial satellite.

A soon-to-launch NASA mission is set to offer more data on North America's air quality than researchers and public health programs have ever had available before. A new monitoring instrument hosted on a commercial satellite will offer continent-wide, hourly updates on air pollution, at about a four square mile resolution, according to the space agency and partner groups behind the project.

The Tropospheric Emissions Monitoring of Pollution (TEMPO) instrument will keep tabs on the air quality in a region stretching from the oil sands in Canada to south of Mexico City, and from the Atlantic to Pacific coasts, according to a NASA news statement.

The device will use advanced imaging to collect detailed light scattering data (i.e. hyperspectral data) on sunlight reflected off of Earth's atmosphere at varying levels. Using this information, scientists back on the ground will be able to translate those wavelengths of light into local concentrations of different pollutants in the air, explained Dennis Nicks, director of payload engineering at Ball Aerospace–the company that NASA contracted to design and build the TEMPO instrument, during a Wednesday press briefing.

upstart writes:

Regulators have to figure out how to get them on the market:

Ella Balasa was 26 when she realized the routine medical treatments that sustained her were no longer working. The slender lab assistant had lived since childhood with the side effects of cystic fibrosis, an inherited disease that turns mucus in the lungs and other organs into a thick, sticky goo that gives pathogens a place to grow. To keep infections under control, she followed a regimen of swallowing and inhaling antibiotics—but by the beginning of 2019, an antibiotic-resistant bacterium lodged in her lungs was making her sicker than she had ever been.

Balasa's lung function was down to 18 percent. She was feverish and too feeble to lift her arms over her head. Even weeks of intravenous colistin, a brutal last-resort antibiotic, made no dent. With nothing to lose, she asked a lab at Yale University whether she could volunteer to receive the organisms they were researching: viruses that attack bacteria, known as bacteriophages.

[...] And it worked. The viruses penetrated the goo, attacked the bacteria, and killed a portion of them; the rest of the bacteria weakened enough that antibiotics could knock them out. Balasa's body cleared the life-threatening infection faster than ever before.

[...] There's an asterisk to her success: Phages are unapproved drugs, not just in the United States, but in the United Kingdom and Western Europe, too. No company makes them for commercial sale in those countries, and hospitals and pharmacies don't stock them. To administer them, physicians must seek a compassionate-use authorization from a government regulator—in Balasa's case, the US Food and Drug Administration—showing their patients have no other options.

An Anonymous Coward writes:

Australia will ban tiktok on government devices despite claims by chinese officials that the application is safe to use.

Is any application on a mobile device really safe to use? What personal data do they collect? Where do they send it? Why don't mobile devices come with the firewall enabled?

Australia's top spy agency has added to growing concerns about a popular social media app, and its collection of users' personal data. State governments across the nation are issuing TikTok bans on official work devices as concerns about data safety increase worldwide.

The app's Australian general manager Lee Hunter recently told The Project that users should feel "safe" on TikTok, and claimed China had no way of accessing data – despite the site's parent company operating out of China.

However, national intelligence organisation Australian Signals Directorate (ASD) recently released advice about the app, warning the general public not to use it on a device that can access other information.

"Do not use it on a phone that can access any official information, for example, any workplace communication (email clients, MS Teams)," the ASD warned in advice shared by the Tasmanian government.

Previously:
    The 'Insanely Broad' RESTRICT Act Could Ban VPNs in the USA
    Banning TikTok
    TikTok Would be Banned From US "for Good" Under Bipartisan Bill
    President Trump Threatens TikTok Ban, Microsoft Considers Buying TikTok's U.S. Operations[Updated 2]

Original Submission

Alcohol Recovery Startups Monument and Tempest Shared Patients' Private Data With Advertisers

upstart writes:

Alcohol recovery startups Monument and Tempest shared patients' private data with advertisers:

For years, online alcohol recovery startups Monument and Tempest were sharing the personal information and health data of their patients with advertisers without their consent.

Monument, which acquired Tempest in 2022, confirmed the extensive years-long leak of patients' information in a data breach notification filed with California's attorney general last week, blaming their use of third-party tracking systems developed by ad giants including Facebook, Google, Microsoft and Pinterest.

In its disclosure, the companies confirmed their use of website trackers, which are small snippets of code that share information about visitors to their websites with tech giants, and often used for analytics and advertising.

The data shared with advertisers includes patient names, dates of birth, email and postal addresses and phone numbers, and membership numbers associated with the companies and patients' insurance provider. The data also included the person's photo, unique digital ID, what services or plan the patient is using, appointment information, and assessment and survey responses submitted by the patient, which includes detailed responses about a person's alcohol consumption and used to determine their course of treatment.

Monument's own website says these survey answers are "protected" and "used only" by its care team.

upstart writes:

Hackers can inject malicious code into a car's electronic network via headlight or taillight wires, fooling the car into thinking the key is present:

As car security has advanced, the world of auto theft has quickly melded with the world of hacking. The advent of high-tech car keys means that hotwiring is out and methods like relay attacks are the new way to gain unauthorized access to a vehicle. Now, however, it seems that attackers have found a new way to entirely bypass the electronic security on modern cars: A method called CAN injection.

[...] The attack relies on a vehicle's CAN bus, the internal computer network that keeps everything running. If you've ever wondered how your car's engine, body control module, and all the little controllers scattered around the car all communicate, CAN bus is the answer. The system is universal in modern cars, and even aftermarket ECU manufacturers now build CAN integration into their products.

The attack method Tindell lays out relies on physical access to the car's CAN bus, meaning an attacker needs to get to the data wires that run through your car. By tapping into these wires, a thief can inject malicious commands into the network — allowing the thief to wake up the car's computer controllers, falsify the presence of the car key, and drive off. [...]

[...] This attack isn't the easiest to pull off, given that it requires a thief to partially disassemble the target car, but it's powerful when done correctly — entirely bypassing the car's key, unlike relay attacks that simply extend the key's radio range. Tindell lists multiple solutions that automakers can implement, most notably the "zero trust" approach — wherein every device, even within a car's internal CAN bus, needs to verify itself during any communication.

Original Submission

upstart writes:

New research offers an intriguing variation on the famous experiment:

A team of physicists has recreated a classic experiment in particle physics with a dimensional twist, by performing it in time instead of just space. In the process, they showcased the utility of a unique material for future experiments.

The researchers redesigned the double-slit experiment, which was first performed in 1801 and demonstrated a curious trait of light: that it can behave both as a particle and a wave.

The original experiment tested how light moved through space; the recent rendition tested how light would travel if it was only allowed to at specific times. Details of the team's work are published today in Nature Physics.

[...] In the traditional experiment, a beam of light is projected at a barrier, behind which is a photosensitive detector. The barrier has two parallel slits. If you threw ordinary matter at the barrier, they would roughly form the same shape as the slits on the detector. But when light is cast on the slits, it splits into two waves that pass through the barrier and intersect on the other side.

[...] In the new work, the researchers made one big change: They swapped out an ordinary screen with two slits for an indium-tin-oxide film—the same material that is used in modern phone screens. (It's actually a metamaterial, or a material not found in nature, that is engineered to behave a certain way.)

The researchers changed the reflectance of the screen using ultrafast lasers, which only allowed light through at specific times, quadrillionths of a second apart. Basically, they built a quantum tollbooth, using the lasers to create time-dependent slits in the barrier. Even when they only allowed one photon through the screen, the light produced an interference pattern.

Journal Reference:
Tirole, Romain, Vezzoli, Stefano, Galiffi, Emanuele, et al. Double-slit time diffraction at optical frequencies, Nature Physics (DOI: 10.1038/s41567-023-01993-w)

Original Submission

RamiK writes:

Bloomberg reports:

Samsung Electronics Co. said it's cutting memory chip production after reporting its slimmest profit since the 2009 financial crisis, a significant step toward ending a supply glut that had cratered prices across the industry.

[...] Operating profit at the world's largest maker of memory chips plunged more than 95% to 600 billion won ($450 million) for the three months ended March, missing the average analyst estimate of 1.4 trillion won. Sales fell 19% to 63 trillion won.

[...] South Korea's largest company had resisted pulling back despite the downturn, in part to grab market share from rivals SK Hynix Inc. and Micron Technology Inc., which have been forced to cut output.

[...] Samsung's shares rose as much as 4.7% in the highest intraday jump in three months. Hynix shares surged as much as 6.7%. [...] The company had warned that earnings would fall in the first quarter on slowing sales. But memory prices tumbled more than anticipated because of sluggish demand for a wide range of electronics from smartphones to PCs, as consumers and companies navigated recession risks. Despite its post-Covid re-opening, China's market has also not bounced back as quickly as some anticipated. Samsung is estimated to have lost about $3 billion in its memory chip division.

[...] Prices of DRAM - a type of memory used to process data - are expected to fall in the current quarter by around 10%, according to Yuanta's Baik. That follows a roughly 20% slide in the previous three months and a more than 30% drop in the fourth quarter of last year. [...] Inventory at Samsung had swelled to 52.2 trillion won at the end of last year after the company maintained production despite a collapse in demand. The existing high inventory will take time to digest, said TrendForce Senior Vice President Avril Wu. "Slower-than-expected demand in the second half of 2023 will prolong the time required for inventory to get back to normal levels."

Original Submission

upstart writes:

Somebody wants to copyright a rhythm – get ready for the dembow tax if they succeed:

One of the most pernicious effects of today's copyright maximalism is the idea that every element of a creative work has to be owned by someone, and protected against "unauthorised" – that is, unpaid – use by other artists. That goes against several thousand years of human creativity, which only exists thanks to successive generations of artists using and building on our cultural heritage. The ownership model of art is essentially selfish: it seeks to maximise the financial gains of one creator, at the expense of the entire culture of which they are part. A good example of this clash of interests can be seen in yet another lawsuit in the music industry. This time, somebody is trying to copyright a rhythm:

The [Fish Market song] track featured the first known example of what would come to be known as a "dembow" rhythm – the percussive, slightly syncopated four-to-the-floor beat that travelled from reggae to become the signature beat of reggaeton, today the world-conquering sound of Latin American pop.

Now, more than 30 years after Fish Market was released, Steely & Clevie Productions is suing three of reggaeton's most celebrated hitmakers – El Chombo, Luis Fonsi and Daddy Yankee – for what they characterise as unlawful interpolation of Fish Market's rhythm (or "riddim"), and are seeking the credit – and royalties – they say they deserved from the start.

upstart writes:

Synthetic embryos made without eggs and sperm are looking increasingly like the real thing—raising thorny ethical questions:

[...] These "synthetic embryos" can be made without the direct contribution of egg or sperm cells. Because they're not "real" embryos, some have argued that the same restrictions don't apply.

Embryos made from stem cells, rather than an egg and sperm, appear to generate a short-lived pregnancy-like response in monkeys.

Recent advances are allowing scientists to create embryo-like structures that look more and more like the real thing. Just this week, scientists in China described how they developed structures called blastoids for 17 days in the lab. They even managed to get some of them to implant in the uteruses of monkeys and trigger the very first signs of pregnancy.

The blastoids didn't survive for very long, probably because researchers haven't quite figured out how best to mimic what happens during the development of a conventional embryo. But most believe that it's just a matter of time. If we can eventually get stem cells to form a viable embryo, a functional fetus, or even a baby, should we treat blastoids in the same way we treat embryos?

Perhaps the bigger question rests on how embryo-like these stem-cell-derived structures are. For some scientists, it's a catch-22 situation. If the blastoids look too much like embryos, then many believe research with them should be restricted in the same way that we control work on human embryos.

But if they don't look enough like embryos, then there's no point in using them for research, says Chuva de Sousa Lopes. "At the moment, it's so difficult to understand how close they are, or how different they are," she says.

upstart writes:

With Zen 5 CPU and RDNA 3+ iGPU, Strix Point could be a game changer:

At CES 2023, AMD announced its Phoenix-HS APUs featuring Zen 4 CPUs and RDNA 3 iGPUs. The company is expected to follow that with the 'Strix Point' APUs, sporting Zen 5 CPUs and RDNA 3+ iGPUs. According to AMD's official roadmap, the Strix Point chips will launch sometime in 2024, although the exact ETA remains a mystery for now.

Almost a year after AMD officially teased its 'Strix Point' APUs, popular YouTube channel RedGamingTech has shared many interesting details about the upcoming hardware. According to the video, the flagship Strix Point APU will come with 16 cores, including 8 big Zen 5 and 8 smaller Zen 4D cores, alongside 32MB of L3 cache.

On the graphics side, the chip is expected to feature an RDNA 3+ iGPU with 8 Work Group Processors (WGP). The video also suggests that the iGPU on the most powerful Strix Point APU would have clock speeds of more than 3GHz.

If the reported specs hold up, it would be a little disappointing, as an earlier video from the channel claimed that Strix Point's RDNA 3+ iGPU could have 12 WGP/24 Compute Units (CU) with around 9 TFLOPS of FP32 precision performance.

Meanwhile, the mid-range Strix Point SKU will reportedly have 4x Zen 5 and 8x Zen 4D cores, paired with 16MB of L3 cache and 4 RDNA 3+ WGPs. Finally, the entry-level chip is said to come with up to 6 cores, including 2x Zen 5 and 4x Zen 4D. They are also tipped to have 8 MB of L3 cache and a 2 WGP iGPU.

The core count, cache and WGPs will not be the only differences between the high-end and entry-level Strix Point APUs. As per the video, the top-end chip in the lineup will be fabricated using TSMC's 3 nm process, while the other two SKUs will use the 4 nm process. The chips are also expected to have a powerful memory controller with support for DDR5-6400 or LPDDR5X-8533 RAM.

Original Submission