SoylentNews

fliptop writes:

The tool could let teachers spot plagiarism or help social media platforms fight disinformation bots:

Hidden patterns purposely buried in AI-generated texts could help identify them as such, allowing us to tell whether the words we're reading are written by a human or not.

These "watermarks" are invisible to the human eye but let computers detect that the text probably comes from an AI system. If embedded in large language models, they could help prevent some of the problems that these models have already caused.

For example, since OpenAI's chatbot ChatGPT was launched in November, students have already started cheating by using it to write essays for them. News website CNET has used ChatGPT to write articles, only to have to issue corrections amid accusations of plagiarism. Building the watermarking approach into such systems before they're released could help address such problems.

In studies, these watermarks have already been used to identify AI-generated text with near certainty. Researchers at the University of Maryland, for example, were able to spot text created by Meta's open-source language model, OPT-6.7B, using a detection algorithm they built. The work is described in a paper that's yet to be peer-reviewed, and the code will be available for free around February 15.

[...] There are limitations to this new method, however. Watermarking only works if it is embedded in the large language model by its creators right from the beginning. Although OpenAI is reputedly working on methods to detect AI-generated text, including watermarks, the research remains highly secretive. The company doesn't tend to give external parties much information about how ChatGPT works or was trained, much less access to tinker with it. OpenAI didn't immediately respond to our request for comment.

Related:

• BuzzFeed Preps AI-Written Content While CNET Fumbles
• Seattle Public Schools Bans ChatGPT; District 'Requires Original Thought and Work From Students'
• ChatGPT Arrives in the Academic World
• OpenAI's New ChatGPT Bot: 10 "Dangerous" Things it's Capable of

Original Submission

upstart writes:

And while Intel is fighting for its life, the rest of the industry is moving on:

In reality, Intel is not the giant of the industry. Intel's total share of industry capacity is around 10%, they are not a giant who has stumbled, they are a niche player and have been for years. Admittedly, they occupy a high-value, high-price niche, but it is a niche nonetheless.

The best analogy we can think of here is automobiles. Mercedes sells around 10% of cars in the US, just as Intel has about 10% of industry capacity. Now imagine if Mercedes somehow lost its brand – maybe a massive recall or a series of high profile vehicle-caused accidents. They would not only lose market share but also all their brand value, causing a long term downward sales trend that would be very expensive to dig their way out. Intel is the luxury brand of semis and suddenly their cars do not move fast. We have tortured that analogy enough, the point is that Intel really does not occupy the strategic high ground we all thought it did.

After their last set of results, especially their guidance for 2023, we are increasingly of the opinion that Intel is out of options. They forecast they are going to burn $15 billion in cash next year, a huge amount even for a company with $34 billion of net cash on their balance sheet.

After their disastrous roadmap event last month, we have to call in to question the company's ability to accurately forecast their business. We actually have many more examples of systematic flaws in their forecasting abilities, but none as public as that event. So we have little confidence in the company's $15 billion forecast, it could easily be much higher. Add to that the need to continue to fund their manufacturing needs and their cash needs are immense.

Nor is it clear if 2024 will be any better. At heart, we have always argued that the company has one task before it and that is an existential task – it has to catch up in manufacturing. The earliest they forecast achieving that is late 2024, which means it will likely not factor into results until 2025. By that time the company's bank balances will be dangerously low.

upstart writes:

The Supreme Court may overhaul how you live online:

On my Google Discover page, for example, I was seeing loads of stories about cancer and grief, which is not in line with the company's targeting policies that are supposed to prevent the system from serving content on sensitive health issues.

Imagine how dangerous it is for uncontrollable, personalized streams of upsetting content to bombard teenagers struggling with an eating disorder or tendencies toward self-harm. Or a woman who recently had a miscarriage, like the friend of one reader who wrote in after my story was published. Or, as in the Gonzalez case, young men who get recruited to join ISIS.

So while the case before the justices may seem largely theoretical, it is really fundamental to our daily lives and the role that the internet plays in society. As Farid told me, "You can say, 'Look, this isn't our problem. The internet is the internet. It reflects the world'... I reject that idea." But recommendation systems organize the internet. Could we really live without them?

Speaking of the State of the Union, Biden called out Big Tech several times, offering the clearest signal yet that there will be increased activity around tech policy—one of the few areas with potential for bipartisan agreement in the newly divided Congress.

upstart writes:

A new crop of unowned seeds is bringing biodiversity back to farming:

Alexander Klepnev CC BY 4.0

When Jack Kloppenburg looks out over his sprawling vegetable garden in rural Wisconsin, he sees half a dozen arm-thick green-striped squash called Candystick Dessert Delicata, and a gaggle of bright yellow Goldini squash among the lush green. "These are so delicious!" he exclaims with all the enthusiasm only a lifelong gardener can muster. But what's special about the vegetables is not just their taste: They have all been grown from open source seeds developed by Oregon farmer Carol Deppe, a Harvard-trained geneticist and board chair of the Open Source Seed Initiative (OSSI).

Most people have heard of open source software, maybe also of open source beer (Free beer for all!) or open source pharmaceutical research. The principle is the same: Someone developed the seeds — for cowpeas, corn, rye and more — and now offers the resource for everybody to share. 

Just like software development has been co-opted by a few global companies like Microsoft and Apple, the international seed development and trade, too, is controlled by a few big giants like Bayer (Monsanto), Corteva (DuPont) and ChemChina (Syngenta). A 2012 Oxfam study found that four companies dominate more than 60 percent of the global trade with grains.

When we buy cereal or bread, few pay attention to the fact that most grains are protected or even patented. Most farmers don't own the seeds they sow on their fields. "They are renting them," Kloppenburg, professor emeritus at the University of Wisconsin-Madison and co-founder of OSSI says with disgust. The problem with that? "A few global companies have the monopolies on global seed trade, and they breed cash crops like corn and soy, purely for money. They don't care about biodiversity, world hunger or about the small farmer." What sounds like a business problem impacts everybody, Kloppenburg insists. "These few gene giants on top of the food chain decide what ends up on our plates."

upstart writes:

Failed Webb calibration leads to discovery of tiny asteroid:

With any new technology, there are bound to be failures — and that's true of cutting-edge astronomy instruments like the James Webb Space Telescope as well. But failures can have a silver lining, as was demonstrated recently when an unsuccessful attempt to calibrate a Webb instrument to a well-known asteroid turned up a delightful surprise: the discovery of a new, different asteroid that is just a few hundred feet across.

Researchers were looking through the data collected during the calibration of Webb's Mid-Infrared Instrument (MIRI) when it was pointed toward known asteroid 1998 BC1 — a procedure that had failed due to technical issues. They hoped that they could use this data to test out some new techniques, but when they went digging, they spotted something unexpected. There was a tiny asteroid around 100 to 200 meters (300 to 650 feet) long that happened to be passing through the instrument's field of view at the same time.

"Our results show that even 'failed' Webb observations can be scientifically useful, if you have the right mindset and a little bit of luck," said lead author of the research, Thomas Müller, in a statement. "Our detection lies in the main asteroid belt, but Webb's incredible sensitivity made it possible to see this roughly 100-meter object at a distance of more than 100 million kilometers."

The smaller targets like asteroids are, the harder they are to detect as they reflect so little light. So it's exciting that Webb was able to detect this new object, thought to be the smallest asteroid Webb has observed so far.

Journal Reference:
T. G. Müller, M. Micheli, T. Santana-Ros, et al. Asteroids seen by JWST-MIRI: Radiometric size, distance, and orbit constraints [open], Astronomy & Astrophysics (DOI: 10.1051/0004-6361/202245304)

Original Submission

upstart writes:

Emperor Penguin promised a relaxed seasonal development cycle and has delivered:

Work on version 6.2 of the Linux kernel will stretch into an eighth release candidate, despite emperor penguin Linus Torvalds now saying it isn't really necessary.

In late January, Torvalds rated release candidate five as "fairly sizable" due to "pent up testing and fixes from people being off."

He therefore flagged his intention to extend this kernel development cycle beyond his preferred seven release candidates.

Release candidate six then landed in what Torvalds described as "suspiciously small" stature, followed by release candidate seven that he labelled "fairly small and controlled, to the point where normally I'd just say that this is the last rc.

"But since I've stated multiple times that I'll do an rc8 due to the holiday start of the release, that's what I'll do."

And indeed he did. On February 12 he announced rc8 and declared: "the only real reason for an rc8 is – as now mentioned several times – just to make up for some time during the holiday season. Not that we seem to really have needed it, but there was also no real reason to deviate from the plan."

Torvalds expects "a few late regression fixes" that will come in handy, but doesn't feel the slightly longer-than-usual development process will prove harmful.

All of which means penguinistas need to wait until next week for additions such as improved performance on many-core IBM POWER servers, persistent memory support for RISC-V, and more enabling code for Intel's "On Demand" software-defined silicon plan.

Original Submission

upstart writes:

Ford's reportedly working on a $3.5 billion battery plant in Michigan:

Ford's gearing up to announce a $3.5 billion battery plant in Marshall, Michigan, according to a report from Automotive News. In an advisory obtained by the outlet, the automaker says it will reveal the news of the factory on Monday in partnership with China's Contemporary Amperex Technology (CATL), a company that creates lithium iron phosphate batteries for electric vehicles (EVs), including the Mustang Mach-E.

While Ford still hasn't confirmed these plans, Michigan reportedly offered $1 billion in incentives to attract the automaker. The factory's expected to bring 2,500 jobs to the area.

The project is part of Ford's efforts to comply with the strict rules set by the Biden administration's Inflation Reduction Act (IRA), which lets EVs assembled in North America qualify for a $7,500 tax credit. Although the IRA also outlines battery sourcing requirements that discourage dealings with "foreign entities of concern," like China, we still don't know how the Treasury Department will interpret these rules until sometime next month.

[...] Even still, Ford's reportedly pushing ahead with the project. It could implement a "novel ownership structure" that would allow the automaker to work with the Chinese company and still qualify for the federal tax credit, according to a report from Bloomberg. This could involve Ford taking 100 percent ownership of the plant itself, while CATL controls operations at the facility and keeps the technology it uses to build the batteries.

Original Submission

upstart writes:

You can now use the $6 boards on-board Bluetooth radio:

The Holy Grail, the Ark of the Covenant, the Crystal Skull, Bluetooth on a Raspberry Pi Pico W. The last of these highly sought after things is finally available, at least to developers. When it launched last June, the $6 Pico W microcontroller impressed everyone with its built-in Wi-Fi 4 , but developers were disappointed that the Bluetooth radio that was built into the board's Infineon CYW43439 wireless chip could not be used.

We've been hearing hints that Bluetooth support was coming for a while now and, as of Friday, the official Raspberry Pi Pico SDK supports it. Available on Raspberry PI's Github repository, SDK 1.5.0 (opens in new tab) adds a new Bluetooth API from BTstack (opens in new tab).

[...] Note that the SDK is made for programming in C or C++ so, if you want to program your Pico W with it, you'll need to use that language. The Github page has instructions (opens in new tab) on how to set up your environment and get started.

We'd prefer to program our Pico W with CircuitPython or MIcroPython, which are much easier to work with. CircuitPython, in particular, has built-in support for turning a Pico or other RP2040-powered microcontroller into an HID device such as a mouse or keyboard. Perhaps now that the official SDK has added Bluetooth support, these other languages will get it also.

Original Submission

upstart writes:

Carmack sees a 60% chance of achieving initial success in AGI by 2030:

[Ed note: This interview is chopped way down to fit here. Lots of good stuff in the full interview. --hubie]

Inside his multimillion-dollar manse on Highland Park's Beverly Drive, Carmack, 52, is working to achieve AGI through his startup Keen Technologies, which raised $20 million in a financing round in August from investors including Austin-based Capital Factory.

This is the "fourth major phase" of his career, Carmack says, following stints in computers and pioneering video games with Mesquite's id Software (founded in 1991), suborbital space rocketry at Mesquite-based Armadillo Aerospace (2000-2013), and virtual reality with Oculus VR, which Facebook (now Meta) acquired for $2 billion in 2014. Carmack stepped away from Oculus' CTO role in late 2019 to become consulting CTO for the VR venture, proclaiming his intention to focus on AGI. He left Meta in December to concentrate full-time on Keen.

Many are predicting stupendous, earth-shattering things will result from this, right?

I'm trying not to use the kind of hyperbole of really grand pronouncements, because I am a nuts-and-bolts person. Even with the rocketry stuff, I wasn't talking about colonizing Mars, I was talking about which bolts I'm using to hold things together. So, I don't want to do a TED talk going on and on about all the things that might be possible with plausibly cost-effective artificial general intelligence.

[...] You'll find people who can wax rhapsodic about the singularity and how everything is going to change with AGI. But if I just look at it and say, if 10 years from now, we have 'universal remote employees' that are artificial general intelligences, run on clouds, and people can just dial up and say, 'I want five Franks today and 10 Amys, and we're going to deploy them on these jobs,' and you could just spin up like you can cloud-access computing resources, if you could cloud-access essentially artificial human resources for things like that—that's the most prosaic, mundane, most banal use of something like this.

If all we're doing is making more human-level capital and applying it to the things that we're already doing today, while you could say, 'I want to make a movie or a comic book or something like that, give me the team that I need to go do that,' and then run it on the cloud—that's kind of my vision for it.

upstart writes:

Intel Says Goodbye to Rocket Lake CPUs:

Intel this week announced plans to discontinue its higher-end 11th Generation Core 'Rocket Lake' processors, which are made using its 14nm-class process technology. The CPUs will still be available to Intel's partners for a while, but their days are now numbered. Intel also said it will phase out its 400 and 500-series chipsets for processors in LGA1200 packaging.

[...] Intel's Rocket Lake processors for desktops have always been somewhat controversial: On one hand, they're based on the Cypress Cove microarchitecture (which derives from Sunny Cove microarchitecture) and are equipped with an Xe-powered integrated GPU, just like Intel's 10nm Ice Lake and Tiger Lake CPUs for mobile PCs and compact desktops. On the other hand — unlike Ice Lake and Tiger Lake processors — Rocket Lake chips are made using a refined 14nm-class process technology.

Because Rocket Lake chips feature backported general-purpose cores, Intel had to reduce the core count of these CPUs from 10 (in the case of Comet Lake) to eight. The new chips still offered better performance than their predecessors in loads of applications, but those who needed higher core count preferred AMD's Ryzen 9 3900-series CPUs with 12 or 16 cores, or even Intel's 10th Generation Comet Lake processors.

It's unlikely that Intel's 11th Generation Core Rocket Lake CPUs will be missed all that much, as Intel has since released two 10nm-based product families for desktops featuring competitive microarchitectures. But for those who would like to upgrade their LGA1200 machines, Rocket Lake chips will continue to be available for a while — but not forever.

Previously:
    AnandTech Reviews Intel's i7-11700K "Rocket Lake" CPU Early
    Gigabyte Confirms Intel Rocket Lake Desktop CPUs Will Launch in March

Original Submission

upstart writes:

The causes of depression are much more complex than the serotonin hypothesis suggests:

You'd be forgiven for thinking that depression has a simple explanation.

The same mantra — that the mood disorder comes from a chemical imbalance in the brain — is repeated in doctors' offices, medical textbooks and pharmaceutical advertisements. Those ads tell us that depression can be eased by tweaking the chemicals that are off-kilter in the brain. The only problem — and it's a big one — is that this explanation isn't true.

The phrase "chemical imbalance" is too vague to be true or false; it doesn't mean much of anything when it comes to the brain and all its complexity. Serotonin, the chemical messenger often tied to depression, is not the one key thing that explains depression. The same goes for other brain chemicals.

The hard truth is that despite decades of sophisticated research, we still don't understand what depression is. There are no clear descriptions of it, and no obvious signs of it in the brain or blood.

The reasons we're in this position are as complex as the disease itself. Commonly used measures of depression, created decades ago, neglect some important symptoms and overemphasize others, particularly among certain groups of people. Even if depression could be measured perfectly, the disorder exists amid myriad levels of complexity, from biological confluences of minuscule molecules in the brain all the way out to the influences of the world at large. Countless combinations of genetics, personality, history and life circumstances may all conspire to create the disorder in any one person. No wonder the science is stuck.

Journal Reference:
Moncrieff, J., Cooper, R.E., Stockmann, T. et al. The serotonin theory of depression: a systematic umbrella review of the evidence. Mol Psychiatry (2022). (DOI: https://doi.org/10.1038/s41380-022-01661-0)

Original Submission

upstart writes:

Google bookended Microsoft's big AI search announcement with underwhelming AI news of its own:

Microsoft has officially taken the lead in the race to build a search engine powered by generative AI. On Tuesday, the company debuted the rumored OpenAI-infused versions of its Bing search engine and Edge web browser, proclaiming them to be the next evolution of the internet — an evolution that so far seems to be crafted by Microsoft. Not Google. And Google seems increasingly concerned about that.

Google has spent the last two decades as the most popular search engine in the world. Search is also Google's biggest revenue stream, thanks to all the ads it places all over search results. So it's unusual to see the company scrambling when it comes to what's always been its bread and butter. Yet, that seems to be exactly what Google is doing in response to Microsoft's plans to integrate AI into Bing, its own search engine, which seem to be further along than Google's. After Microsoft invited journalists to see its new AI products last week, Google scrambled to make announcements and show off demos of its own. In other words, Google, which long since surpassed Microsoft's search and web browsing tools, is now playing catch-up.

[...] Bard and the new Bing seem pretty similar on the surface. But it's hard to say without trying them, and neither is available to the general public yet. They're both rolling out in the next few weeks. But while Bard is built on a "lightweight" version of its generative chatbot for now, Microsoft says the new Bing will use an even more powerful version of ChatGPT that was custom-designed for search. And while Bard was introduced in a short blog post, Microsoft invited a ton of journalists to a splashy live event at the company's headquarters in Redmond, Washington, to show off its AI-powered Bing and Edge. This all suggests that one company thinks its AI search is ready for primetime, while the other is just trying not to be left out of the conversation completely.

upstart writes:

When it comes to evaluating COVID treatments, MDs are only human:

[...] We'd like to think people like doctors would carefully evaluate evidence before making treatment decisions, yet a correlation between voting patterns and ivermectin prescriptions suggests that they don't.

Of course, a correlation at that sort of population level leaves a lot of unanswered questions about what's going on. A study this week tries to fill in some of those blanks by performing controlled experiments with a set of MDs. The work clearly shows how ideology clouds professional judgments even when it comes to reading the results of a scientific study.

The work primarily focuses on a panel of about 600 critical care physicians—the people who are most likely to be the first source of treatment for those who develop severe COVID-19. It also involved a panel of 900 people who aren't involved in medicine to provide a comparison population. While some initial surveys were done earlier, most of the data comes from the spring of 2022, long after COVID-19 vaccines had established their effectiveness in limiting severe symptoms of the disease. By then, a couple of widely hyped "cures"—hydroxychloroquine and ivermectin—had been definitively ruled out as therapeutic.

All the participants were asked to self-rate on a seven-point scale, from very liberal to very conservative. For most studies, the answers from the liberal and conservative participants were evaluated in terms of how greatly they differed from those of the moderate participants.

fliptop writes:

More than 80 years ago, science fiction writer Isaac Asimov wrote Reason, a story about a solar energy collection station in deep space that delivered high-energy rays to receivers on Earth and Mars. Called space-based solar power (SSP), Asimov's idea didn't start to approach science fact in any meaningful sense until 1968...

...when Peter Glaser, an aerospace engineer with Apollo program experience working for the Arthur D. Little consultancy in Cambridge, MA, published a paper suggesting ways to construct SSP stations with separate solar collecting and giant dish-based microwave transmitters.

However, based on the pre-carbon-fiber, heavy-metal aerospace technologies of the day, studies by Nasa and the U.S. Department of Energy determined that a single solar-receiving satellite would weigh in excess of 80,000 tons, putting launch costs per power station way beyond consideration.

On January 3 2023, a SpaceX Falcon 9 rocket launched 114 small payloads, and right now...

...in a sun-synchronous orbit about 525 kilometers overhead, there is a small experimental satellite called the Space Solar Power Demonstrator One (SSPD-1 for short). It was designed and built by a team at the California Institute of Technology, funded by donations from the California real estate developer Donald Bren[.]

"To the best of our knowledge, this would be the first demonstration of actual power transfer in space, of wireless power transfer," says Ali Hajimiri, a professor of electrical engineering at Caltech and a codirector of the program behind SSPD-1, the Space Solar Power Project.

The Caltech team is waiting for a go-ahead from the operators of a small space tug to which it is attached, providing guidance and attitude control. If all goes well, SSPD-1 will spend at least five to six months testing prototype components of possible future solar stations in space.

[...] If it works out, in 30 years maybe there could be orbiting solar power fleets, adding to the world's energy mix. In other words, as a recent report from [British engineering consultancy] Frazer-Nash concluded, this is "a potential game changer."

SSPD-1 story originally spotted on The Eponymous Pickle.

Previously: Space-Based Solar Power Hardware Ready for Actual Testing in Space

Original Submission

On the 12 February 2014, SoylentNews published its first 'official' discussion wittily entitled "Welcome to SoylentNews!"

There were 23 comments to that story but that belies the incredible effort made by a small team of enthusiasts who were determined to create and manage a site that lived up to their own expectations of how a site should behave. There are no advertisements and we are not beholden to any corporate group or business interest. All the support is still completely voluntary. We have also committed ourselves to reaching at least the 10 year birthday celebration too.

In the first story, a certain NCommander made the following remark:

Now that we're here, we hope to have made the wait worth it, but we depend on everyone in the community. To make this site a success, we depend on each and every single user even if its just from passing word of mouth. Remember, every single user can submit stories, moderate, and contribute to discussions all at the same time, and that's what makes us unique. May I be the first to welcome you to your new home.