SoylentNews

Freeman writes:

https://arstechnica.com/information-technology/2024/09/ai-superintelligence-looms-in-sam-altmans-new-essay-on-the-intelligence-age/

On Monday, OpenAI CEO Sam Altman outlined his vision for an AI-driven future of tech progress and global prosperity in a new personal blog post titled "The Intelligence Age." The essay paints a picture of human advancement accelerated by AI, with Altman suggesting that superintelligent AI could emerge within the next decade.

"It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I'm confident we'll get there," he wrote.

OpenAI's current goal is to create AGI (artificial general intelligence), which is a term for hypothetical technology that could match human intelligence in performing many tasks without the need for specific training. By contrast, superintelligence surpasses AGI, and it could be seen as a hypothetical level of machine intelligence that can dramatically outperform humans at any intellectual task, perhaps even to an unfathomable degree.
[...]
Despite the criticism, it's notable when the CEO of what is probably the defining AI company of the moment makes a broad prediction about future capabilities—even if that means he's perpetually trying to raise money. Building infrastructure to power AI services is foremost on many tech CEOs' minds these days.

"If we want to put AI into the hands of as many people as possible," Altman writes in his essay, "we need to drive down the cost of compute and make it abundant (which requires lots of energy and chips). If we don't build enough infrastructure, AI will be a very limited resource that wars get fought over and that becomes mostly a tool for rich people."
[...]
While enthusiastic about AI's potential, Altman urges caution, too, but vaguely. He writes, "We need to act wisely but with conviction. The dawn of the Intelligence Age is a momentous development with very complex and extremely high-stakes challenges. It will not be an entirely positive story, but the upside is so tremendous that we owe it to ourselves, and the future, to figure out how to navigate the risks in front of us."
[...]
"Many of the jobs we do today would have looked like trifling wastes of time to people a few hundred years ago, but nobody is looking back at the past, wishing they were a lamplighter," he wrote. "If a lamplighter could see the world today, he would think the prosperity all around him was unimaginable. And if we could fast-forward a hundred years from today, the prosperity all around us would feel just as unimaginable."

Related Stories on Soylent News:
Plan Would Power New Microsoft AI Data Center From Pa.'s Three Mile Island 'Unit 1' Nuclear Reactor - 20240921
Artificial Intelligence 'Godfather' on AI Possibly Wiping Out Humanity: 'It's Not Inconceivable' - 20230329
Microsoft Research Paper Claims Sparks of Artificial Intelligence in GPT-4 - 20230327
John Carmack's 'Different Path' to Artificial General Intelligence - 20230213

Original Submission

upstart writes:

Just in time for spooky season:

If you're thinking of going down the unconventional route for this year's Halloween costume, we've got just the inspiration for you: a brand-new species of chimaera, better known as a ghost shark or spookfish.

Scientists discovered the new species, which goes by the scientific name Harriotta avia – avia meaning "grandmother" in Latin, in honor of study author Dr Brit Finucci's grandmother – and the common name Australasian Narrow-nosed Spookfish, off the coast of New Zealand and Australia.

This region is no stranger to ghost sharks – the study describing the new species dubs it "a global diversity [hotspot] for chimaeroids" – but it was previously thought that this particular species wasn't actually its own species at all. Instead, it was believed to be a variant of another species, Harriotta raleighana, which is found across the globe.

But with the help of genetics and a closer inspection of its morphology, the researchers were able to identify H. avia as a species in its own right.

"Harriotta avia is unique due to its elongated, narrow and depressed snout; long, slender trunk; large eyes; and very long, broad pectoral fins," said Finucci in a statement. "It is a lovely chocolate brown colour."

[...] "Ghost sharks like this one are largely confined to the ocean floor, living in depths of up to 2,600 [meters] [8530 feet]," said Finucci. The level of pressure to be found at such depths isn't exactly human-friendly.

"Their habitat makes them hard to study and monitor, meaning we don't know a lot about their biology or threat status, but it makes discoveries like this even more exciting," Finucci explained.

What we do know is that, despite the nickname, ghost sharks aren't actually sharks at all. Whilst the two groups are still related to each other, it's thought that they diverged from one another other nearly 400 million years ago. Both remained cartilaginous, but ghost sharks have wound up with several physical differences from their relatives.

Journal Reference:
Finucci, Brittany, Didier, Dominique, Ebert, David A., et al. Harriotta avia sp. nov. – a new rhinochimaerid (Chimaeriformes: Rhinochimaeridae) described from the Southwest Pacific, Environmental Biology of Fishes (DOI: 10.1007/s10641-024-01577-4)

Original Submission

Motor Trend is running a piece on the systems in the recently released Mercedes-Benz "Drive Pilot 95", https://www.motortrend.com/reviews/mercedes-benz-drive-pilot-95-first-drive-review

Here are a few of the details I found interesting:

By the end of this year, pending final certification from the authorities, German customers will be able to order an upgraded version called Drive Pilot 95, which, under certain operating conditions, will allow their S-Class and EQS models to self-drive for an indefinite period in the right lane of autobahns at speeds of up to 95 km/h (59 mph).[The earlier version from 2022 only worked in congested traffic up to 65 kph (40mph)]
[...]
Why has it taken so long to implement a software tweak? Well, both the German legislators and Mercedes-Benz, which assumes legal responsibility for the functioning of its vehicles while they are operating in Level 3 autonomous drive mode, are cautious. The Silicon Valley 'move fast and break things' approach doesn't work for them.
[...]
In addition to the parking sensors in the front and rear bumpers and the 360 degree cameras in the rear view mirrors that are fitted to many Mercedes-Benz models, Drive Pilot equipped cars have multi-mode radars at each corner, a front-facing long-range radar and a lidar unit behind the grille, a stereo camera at the top of the windshield, a regular camera facing rearward through the backlight, and a moisture sensor in the front wheel well.
The rear-facing camera is used to detect the flashing lights of emergency vehicles approaching from behind, though the 'Hey Mercedes' voice activation microphone in the cabin will pick up the sound of the sirens even if the vehicle cannot be seen. The moisture detector, which measures the sound level of the spray from the tire on wet roads, is used to determine whether rain and spray could interfere with the camera, radar and the lidar systems.
[...]
In simple terms, the key difference between the original Drive Pilot system and Drive Pilot 95 is the latter will now operate autonomously at Level 3 for an indefinite period if the Mercedes-Benz is in the right lane of the autobahn and is following traffic traveling at no more than 95km/h. Without that traffic, which can be up to 1000 feet ahead, the system will not activate.

owl writes:

http://www.righto.com/2024/09/ramtron-ferroelectric-fram-die.html

Ferroelectric memory (FRAM) is an interesting storage technique that stores bits in a special "ferroelectric" material. Ferroelectric memory is nonvolatile like flash memory, able to hold its data for decades. But, unlike flash, ferroelectric memory can write data rapidly. Moreover, FRAM is much more durable than flash and can be be written trillions of times. With these advantages, you might wonder why FRAM isn't more popular. The problem is that FRAM is much more expensive than flash, so it is only used in niche applications.

[...] The history of ferroelectric memory dates back to the early 1950s.3 Many companies worked on FRAM from the 1950s to the 1970s, including Bell Labs, IBM, RCA, and Ford. The 1955 photo below shows a 256-bit ferroelectric memory built by Bell Labs. Unfortunately, ferroelectric memory had many problems,4 limiting it to specialized applications, and development was mostly abandoned by the 1970s.

Ferroelectric memory had a second chance, though. A major proponent of ferroelectric memory was George Rohrer, who started working on ferroelectric memory in 1968. He formed a memory company, Technovation, which was unsuccessful, and then cofounded Ramtron in 1984.5 Ramtron produced a tiny 256-bit memory chip in 1988, followed by much larger memories in the 1990s.

Original Submission

Arthur T Knackerbracket has processed the following story:

Lasso peptides are natural products made by bacteria. Their unusual lasso shape endows them with remarkable stability, protecting them from extreme conditions. In a new study, published in Nature Chemical Biology, researchers have constructed and tested models for how these peptides are made and demonstrated how this information might be used to advance lasso peptide-based drugs into the clinic.

"Lasso peptides are interesting because they are basically linear molecules that have been tied into a slip knot-like shape," said Susanna Barrett, a graduate student in the Mitchell lab (MMG). "Due to their incredible stability and engineerability, they have a lot of potential as therapeutics. They have also been shown to have antibacterial, antiviral, and anti-cancer properties."lasso peptides

Lasso peptides are ribosomally synthesized and post-translationally modified molecules. The peptide chains are formed from joining amino acids together in the form of a string, which is done by the ribosome. Two enzymes, a peptidase and a cyclase, then collaborate to convert a linear precursor peptide into the distinctive knotted lasso structure. Since their discovery over three decades ago, scientists have been trying to understand how the cyclase folds the lasso peptide.

"One of the major challenges of solving this problem has been that the enzymes are difficult to work with. They are generally insoluble or inactive when you attempt to purify them," Barrett said.

One rare counterexample is fusilassin cyclase, or FusC, which the Mitchell lab characterized in 2019. Former group members were able to purify the enzyme, and since then, it has served as a model to understand the lasso knot-tying process. Yet, the structure of FusC remained unknown, making it impossible to understand how the cyclase interacts with the peptide to fold the knot.

In the current study, the group used the artificial intelligence program AlphaFold to predict the FusC protein structure. They used the structure and other artificial intelligence-based tools, like RODEO, to pinpoint which cyclase active site residues were important for interacting with the lasso peptide substrate.

"FusC is made up of approximately 600 amino acids and the active site contains 120. These programs were instrumental to our project because they allowed us to do 'structural studies' and whittle down which amino acids are important in the active site of the enzyme," Barrett said.

They also used molecular dynamics simulations to computationally understand how the lasso is folded by the cyclase. "Thanks to the computing power of Folding@home, we were able to collect extensive simulation data to visualize the interactions at the atomic level," said Song Yin, a graduate student in the Shukla lab. "Before this study, there were no MD simulations of the interactions between lasso peptides and cyclases, and we think this approach will be applicable to many other peptide engineering studies."

From their computational efforts, the researchers found that among different cyclases, the backwall region of the active site seemed to be especially important for folding. In FusC, this corresponded to the helix 11 region. The researchers then carried out cell-free biosynthesis where they added all the cell components that are necessary for the synthesis of the lasso peptides to a test tube with enzyme variants that had different amino acids in the helix 11 region. Ultimately, they identified a version of FusC with a mutation on helix 11 that could fold lasso peptides which cannot be made by the original cyclase. This data confirms the model for lasso peptide folding that the researchers developed with their computational approaches.

fliptop writes:

Investors sold after the investment bank's analysts warned about what they called the 'China butterfly effect':

Shares of General Motors and Ford Motor traded lower on Wednesday after Morgan Stanley downgraded the overall U.S. auto sector, citing worries that Western automakers might struggle in the intensifying competition with Chinese rivals.

General Motors was downgraded to "underweight" from "equal weight," and its shares fell 5.4 percentage points, to $45.50. Ford went to "equal weight" from "overweight," with its shares dropping more than 4 percentage points, to $10.43.

Electric vehicle (EV) maker Rivian Automotive and Canadian parts manufacturer Magna International were both downgraded to "equal weight" from "overweight." Shares of Rivian were down 5.7 percentage points while Magna's were off 4.7 percentage points.

Investors sold after Morgan Stanley analysts warned about what they called the "China butterfly effect," a metaphor suggesting that even small surges in China's industrial production capacity could have significant ripple effects across the global market.

[...] Bolstered in part by massive government subsidies, Chinese manufacturers have rapidly emerged as major players in the EV industry, accounting for 60 percent of worldwide EV sales and almost one in five EVs sold in Europe last year.

Both Washington and Brussels have hiked tariffs in response to China's excess production of low-price EVs.

Previously:

• VW Turns on Germany as China Targets Europe's EV Blunders
• Auto Woes
• Chinese Automaker Zotye Plans 2020 Entrance into the U.S. Market

Original Submission

Arthur T Knackerbracket has processed the following story:

UK government IT contracts worth £23.4 billion are due to end during the current five-year Parliament, according to researchers who warn that poor performing suppliers are hardly ever excluded from bidding again.

A report by public spending research company Tussell and the Institute for Government found that a third of these, worth £9 billion, are supposed to finish up in 2025.

The report points out that large contracts expiring next year include the longstanding Post Office deal with Fujitsu to build and manage the Horizon IT system at the center of one of the greatest miscarriages of justice in the UK. From 1999 until 2015, 736 local branch managers were wrongfully convicted of fraud when errors in the system were to blame. The total value of the Horizon contract is £2.38 billion ($3.15 billion). It is due to expire on March 31, 2025.

[...] The researchers warn that poor-performing suppliers to UK government are virtually never excluded from supplying the public sector and often continue to receive government money. Meanwhile, a large number of contracts, totaling billions of pounds, are overseen by officials who are not commercial specialists.

The report also highlights that poor data across government departments meant officials didn't know how much they were spending and with whom. And new providers that could perhaps deliver better services for less money are discouraged from bidding for business.

[...] "Public procurement is a huge market hiding in plain sight, accounting for approximately one-third of all public spending and 10 percent of UK GDP," said Gus Tugendhat, founder of Tussell.

"In the context of tight budgets and strained public services, getting value for money out of government contracts is more important than ever," he said.

Original Submission

looorg writes:

There is a fair, and long running, amount of research for that playing Tetris helps people deal with trauma, PTSD of some kind. Adding some more recent research then where it can reduce PTSD symptoms in healthcare workers (nurses) that worked with trauma COVID19 patients.

Playing something such as Tetris (it's a bit unclear if it's just Tetris or a similar style of games of which Tetris is the prime example) can induce some relaxing zen like state or a "cognitive vaccine". 20 minutes is apparently the prescribed dosage of rotational healing experience. There was the 15 minutes of talking to before playing Tetris. But clearly the healing power of Tetris at work ...

The study was carried out with healthcare workers in Sweden who worked with COVID-19 patients and were exposed to work-related trauma. It was conducted during the COVID-19 pandemic between September 2020 and April 2022. A total of 164 participants were included. Participants were recruited through information at workplaces. Participation was entirely voluntary. The criterion for participation was that the person had at least two intrusive memories per week due to traumatic events that occurred at work.

https://www.uu.se/en/press/press-releases/2024/2024-09-20-ptsd-symptoms-can-be-reduced-through-treatment-including-a-video-game
https://www.psych.ox.ac.uk/news/tetris-used-to-prevent-post-traumatic-stress-symptoms

Original Submission

Arthur T Knackerbracket has processed the following story:

A truck full of lithium-ion batteries is burning in Los Angeles, shutting down ports and a bridge. It’s not clear what the batteries were for — but LA’s Vincent Thomas Bridge, leading to the Port of Los Angeles and the next-door Port of Long Beach, has been shut down for at least 15 hours now while local firefighters let the truck burn. State Route 47 was also closed in both directions as of a couple of hours ago.

Amazingly, a local towing company caught the explosion on camera from a nearby drone:

Both the Port of Los Angeles and the Port of Long Beach have shut down a number of terminals while the fire continues to burn. As of 12:10PM PT on Friday, the truck was still on fire, and both the ports and bridge were still closed, Los Angeles Fire Department (LAFD) spokesperson Ren Medina told The Verge.

Firefighters are nearby and are actively monitoring the situation; as of 10PM PT on Thursday, the fire was expected to last “at least another 24-48 hours.”

As we’ve seen with several EV battery fires, big concentrated lithium battery fires can be very difficult to put out: firefighters sometimes douse them with thousands of gallons of water only to see the fire restart as additional battery cells heat up to the point that they combust. Once a cell gets hot enough, it’s said to go into “thermal runaway,” at which point it can sometimes restart a fire. The LAFD confirms this is a case of thermal runaway.

EV packs are particularly dense with cells, but we don’t yet know if they were involved here — the LA Fire Department spokesperson says it’s not clear who owns the truck, let alone what it was carrying. The LAFD could only confirm they are lithium-ion batteries at this point.

Original Submission

oaklandwatch writes:

A senior RedMonk analyst tried to prove shifting to proprietary licenses *doesn't* improve financial outcomes. But what's interesting is the reactions she got -- from a VC at OSS Capital, ex-Googlers, Chef's co-founder, and even Taylor Dolezal, head of ecosystem at the Cloud Native Computing Community. Plus analyst Lawrence Hecht, who concluded "these companies are nowhere closer to being profitable than before."

There's new quotes from the analyst herself. ("I asked Stephens if she thought the analysis would have an impact in the future on companies considering moves to proprietary licensing. 'I doubt it,' Stephens replied...") And Hecht pounds away at the missteps. ("The assumption has been that closing a company's license will allow the companies to increase their margins among their existing customers... The percentage of companies using a given technology is not changing... Elasticsearch fell from 14% to 13%..")

Interestingly, the study hits right as Elastic is switching *back* to an open-source license. They weigh in in this article too...

It's the discussion about open source licensing that really needed to happen.

Original Submission

Freeman writes:

https://arstechnica.com/information-technology/2024/09/openai-plans-tectonic-shift-from-nonprofit-to-for-profit-giving-altman-equity/

On Wednesday, Reuters reported that OpenAI is working on a plan to restructure its core business into a for-profit benefit corporation, moving away from control by its nonprofit board. The shift marks a dramatic change for the AI company behind ChatGPT, potentially making it more attractive to investors while raising questions about its commitment to sharing the benefits of advanced AI with "all of humanity," as written in its charter.

A for-profit benefit corporation is a legal structure that allows companies to pursue both financial profits and social or environmental goals, ostensibly balancing shareholder interests with a broader mission to benefit society. It's an approach taken by some of OpenAI's competitors, such as Anthropic and Elon Musk's xAI.

[...] Bloomberg reports that OpenAI is discussing giving Altman a 7 percent stake, though the exact details are still under negotiation. This represents a departure from Altman's previous stance of not taking equity in the company, which he had maintained was in line with OpenAI's mission to benefit humanity rather than individuals.

[...] The proposed restructuring also aims to remove the cap on returns for investors, potentially making OpenAI more appealing to venture capitalists and other financial backers. Microsoft, which has invested billions in OpenAI, stands to benefit from this change, as it could see increased returns on its investment if OpenAI's value continues to rise.

DannyB writes:

Linux boots in 4.76 days on the Intel 4004

Historic 4-bit microprocessor from 1971 can execute Linux commands over days or weeks.

Hardware hacker Dmitry Grinberg recently achieved what might sound impossible: booting Linux on the Intel 4004, the world's first commercial microprocessor. With just 2,300 transistors and an original clock speed of 740 kHz, the 1971 CPU is incredibly primitive by modern standards. And it's slow—it takes about 4.76 days for the Linux kernel to boot.

Initially designed for a Japanese calculator called the Busicom 141-PF, the 4-bit 4004 found limited use in commercial products of the 1970s [...]

[....] If you're skeptical that this feat is possible with a raw 4004, you're right: The 4004 itself is far too limited to run Linux directly. Instead, Grinberg created a solution that is equally impressive: an emulator that runs on the 4004 and emulates a MIPS R3000 processor—the architecture used in the DECstation 2100 workstation that Linux was originally ported to.

If it can run a C compiler, it can probably run DOOM.

See Also:

• the YouTube video

Original Submission

Arthur T Knackerbracket has processed the following story:

by University of Texas at Dallas

In a study published July 29 in Advanced Materials, University of Texas at Dallas researchers found that X-rays of the kidneys using gold nanoparticles as a contrast agent might be more accurate in detecting kidney disease than standard laboratory blood tests. Based on their study in mice, they also found that caution may be warranted in employing renal-clearable nanomedicines to patients with compromised kidneys.

Before administering renal-clearable drugs, doctors routinely check a patient's kidney function by testing their blood urea nitrogen (BUN) and creatinine (Cr) levels. With the increasing use of engineered nanoparticles to deliver payloads of drugs or imaging agents to the body, an important question is how the nanoparticles' movement and elimination through the kidney is affected by kidney damage. Can traditional biomarkers like BUN and Cr accurately predict how well—or how poorly—such nanoparticles will move through the kidneys?

The UT Dallas researchers found that in mice with severely injured kidneys caused by the drug cisplatin, in which BUN and Cr levels were 10 times normal, nanoparticle transport through the kidneys was slowed down significantly, a situation that caused the nanoparticles to stay in the kidneys longer.

In mildly injured kidneys, however, in which BUN and Cr levels were only four to five times higher than normal, the transport and retention of gold nanoparticles couldn't be predicted by those tests.

On the other hand, the amount of gold nanoparticle accumulation seen on X-rays did correlate strongly with the degree of kidney damage.

"While our findings emphasize the need for caution when using these advanced treatments in patients with compromised kidneys, they also highlight the potential of gold nanoparticles as a noninvasive way to assess kidney injuries using X-ray imaging or other techniques that correlate with gold accumulation in the kidneys," said Dr. Mengxiao Yu, a corresponding author of the study and a research associate professor of chemistry and biochemistry in the School of Natural Sciences and Mathematics.

Chemistry and biochemistry research scientist Xuhui Ning BS'14, Ph.D.'19 is lead author of the study, and Dr. Jie Zheng, professor of chemistry and biochemistry and a Distinguished Chair in Natural Sciences and Mathematics, is a corresponding author. Other contributors are affiliated with UT Southwestern Medical Center and Vanderbilt University Medical Center.

More information: Xuhui Ning et al, Gold Nanoparticle Transport in the Injured Kidneys with Elevated Renal Function Biomarkers, Advanced Materials (2024). DOI: 10.1002/adma.202402479

Journal information: Advanced Materials

Original Submission

Freeman writes:

https://arstechnica.com/security/2024/09/false-memories-planted-in-chatgpt-give-hacker-persistent-exfiltration-channel/

When security researcher Johann Rehberger recently reported a vulnerability in ChatGPT that allowed attackers to store false information and malicious instructions in a user's long-term memory settings, OpenAI summarily closed the inquiry, labeling the flaw a safety issue, not, technically speaking, a security concern.

So Rehberger did what all good researchers do: He created a proof-of-concept exploit that used the vulnerability to exfiltrate all user input in perpetuity. OpenAI engineers took notice and issued a partial fix earlier this month.

The vulnerability abused long-term conversation memory, a feature OpenAI began testing in February and made more broadly available in September.

[...] Within three months of the rollout, Rehberger found that memories could be created and permanently stored through indirect prompt injection, an AI exploit that causes an LLM to follow instructions from untrusted content such as emails, blog posts, or documents. The researcher demonstrated how he could trick ChatGPT into believing a targeted user was 102 years old, lived in the Matrix, and insisted Earth was flat and the LLM would incorporate that information to steer all future conversations.

[...] The attack isn't possible through the ChatGPT web interface, thanks to an API OpenAI rolled out last year.

[...] OpenAI provides guidance here for managing the memory tool and specific memories stored in it. Company representatives didn't respond to an email asking about its efforts to prevent other hacks that plant false memories.

Original Submission

Arthur T Knackerbracket has processed the following story:

Two of the world's satellite positioning service constellations reached important milestones this week, after the European Space Agency and China's Satellite Navigation Office each launched its own pair of satellites.

Europe's sats were carried by a SpaceX Falcon 9 rocket that left Florida's Kennedy Space Center on September 18. A day later, China's birds rode a Long March 3B that launched from the Xichang Satellite Launch Center in Sichuan province.

China's sats were the 63rd and 64th members of its Beidou constellation, which currently has 50 operating satellites.

This pair were the last of China's third-generation navigation-sat design. Local media reported that the two satellites are spares in case others falter, and that they include some tech that is expected to be included in fourth-gen sats.

[...] Europe's launch delivered the 31st and 32nd members of its Galileo constellation into space.

"With the deployment of these two satellites, Galileo completes its constellation as designed, reaching the required operational satellites plus one spare per orbital plane," proclaimed ESA director of navigation Javier Benedicto.

Original Submission