New Book Warns CS Mindset and VC Industry are Ignoring Competing Values

So apparently three Stanford professors are offering some tough-love to young people in the tech community. Mehran Sahami first worked at Google when it was still a startup (recruited to the company by Sergey Brin). Currently a Stanford CS professor, Sahami explained in 2019 that “I want students who engage in the endeavor of building technology to think more broadly about what are the implications of the things that they’re developing — how do they impact other people? I think we’ll all be better off.”

Now Sahami has teamed up with two more Stanford professors to write a book calling for “a mature reckoning with the realization that the powerful technologies dominating our lives encode within them a set of values that we had no role in choosing and that we often do not even see…”

At a virtual event at Silicon Valley’s Computer History Museum, the three professors discussed their new book, System Error: Where Big Tech Went Wrong and How We Can Reboot — and thoughtfully and succinctly distilled their basic argument. “The System Error that we’re describing is a function of an optimization mindset that is embedded in computer science, and that’s embedded in technology,” says political scientist Jeremy Weinstein (one of the book’s co-authors). “This mindset basically ignores the competing values that need to be ‘refereed’ as new products are designed. It’s also embedded in the structure of the venture capital industry that’s driving the growth of Silicon Valley and the growth of these companies, that prioritizes scale before we even understand anything about the impacts of technology in society. And of course it reflects the path that’s been paved for these tech companies to market dominance by a government that’s largely been in retreat from exercising any oversight.”

Sahami thinks our technological landscape should have a protective infrastructure like the one regulating our roads and highways. “It’s not a free-for all where the ultimate policy is ‘If you were worried about driving safely then don’t drive.'” Instead there’s lanes and traffic lights and speed bumps — an entire safe-driving infrastructure which arrived through regulation.” Or (as their political science professor/co-author Rob Reich tells the site), “Massive system problems should not be framed as choices that can be made by individual consumers.”

Sahami also thinks breaking up big tech monopolies would just leaves smaller “less equipped” companies to deal with the same problems — but that positive changes in behavior might instead come from government scrutiny. But Reich also wants to see professional ethics (like the kind that are well-established in biomedical fields). “In the book we point the way forward on a number of different fronts about how to accelerate that…”

And he argues that at colleges, just one computing-ethics class isn’t enough. “Ethics must be embedded through the entire curriculum.”

Read more of this story at Slashdot.

Trump’s Truth App Bans Criticism of Itself – and Also ‘Excessive Use of Capital Letters’

Time magazine spotted three things in the terms of service for former U.S. president Trump’s “Truth Social” site:

– Despite advertising itself as a platform that will “give a voice to all,” according to a press release, TRUTH Social’s terms of service state that users may not “disparage, tarnish, or otherwise harm, in our opinion, us and/or the Site.” In other words, any user who criticizes Trump or the site can be kicked off the platform…

– [W]hile portraying itself as a refuge for free speech and the “first major rival to ‘Big Tech,'” TRUTH Social’s terms of service make it clear that the platform not only intends to moderate content — just as Twitter and Facebook do — but reserves the right to remove users for any reason it deems necessary. The terms go on to say that if TRUTH Social decides to terminate or suspend your account, the platform may also sue you — something that Twitter and Facebook’s terms don’t say. “In addition to terminating or suspending your account, we reserve the right to take appropriate legal action, including without limitation pursuing civil, criminal, and injunctive redress,” TRUTH Social’s terms state…

– Maybe most notably, the site’s list of prohibited activities includes the “excessive use of capital letters,” an idiosyncrasy that Trump became known for on Twitter and that no other major social network specifically bans. TRUTH Social’s terms also contain some sections written in all-caps.

The terms also specify explicitly that the site considers itself “not responsible” for the accuracy/reliability of what’s posted on the site. Yet the Washington Post reports the newly-formed “Trump Media & Technology Group” has already applied for trademark rights for the terms “truthing,” “post a truth,” and “retruth.”

Meanwhile, the Software Freedom Conservancy believes the end of the site’s public test launch was directly tied to a recently-discovered violation of a Conservancy license. “Once caught in the act, Trump’s Group scrambled and took the site down.”
One of the license’s authors emphasizes that the license “purposefully treats everyone equally (even people we don’t like or agree with), but they must operate under the same rules of the copyleft licenses that apply to everyone else…”

To comply with this important FOSS license, Trump’s Group needs to immediately make that Corresponding Source available to all who used the site today while it was live. If they fail to do this within 30 days, their rights and permissions in the software are automatically and permanently terminated. That’s how AGPLv3’s cure provision works — no exceptions — even if you’re a real estate mogul, reality television star, or even a former POTUS.”

Read more of this story at Slashdot.

DDR4 Memory Protections Are Broken Wide Open By New Rowhammer Technique

“An unprivileged application can corrupt data in memory by accessing ‘hammering’ rows of DDR4 memory in certain patterns millions of times a second, giving those untrusted applications nearly unfettered system privileges,” writes long-time Slashdot reader shoor. Ars Technica reports: Rowhammer attacks work by accessing — or hammering — physical rows inside vulnerable chips millions of times per second in ways that cause bits in neighboring rows to flip, meaning 1s turn to 0s and vice versa. Researchers have shown the attacks can be used to give untrusted applications nearly unfettered system privileges, bypass security sandboxes designed to keep malicious code from accessing sensitive operating system resources, and root or infect Android devices, among other things. All previous Rowhammer attacks have hammered rows with uniform patterns, such as single-sided, double-sided, or n-sided. In all three cases, these “aggressor” rows — meaning those that cause bitflips in nearby “victim” rows — are accessed the same number of times.

Research published on Monday presented a new Rowhammer technique. It uses non-uniform patterns that access two or more aggressor rows with different frequencies. The result: all 40 of the randomly selected DIMMs in a test pool experienced bitflips, up from 13 out of 42 chips tested in previous work (PDF) from the same researchers. “We found that by creating special memory access patterns we can bypass all mitigations that are deployed inside DRAM,” Kaveh Razavi and Patrick Jattke, two of the research authors, wrote in an email. “This increases the number of devices that can potentially be hacked with known attacks to 80 percent, according to our analysis. These issues cannot be patched due to their hardware nature and will remain with us for many years to come.”

The non-uniform patterns work against Target Row Refresh. Abbreviated as TRR, the mitigation works differently from vendor to vendor but generally tracks the number of times a row is accessed and recharges neighboring victim rows when there are signs of abuse. The neutering of this defense puts further pressure on chipmakers to mitigate a class of attacks that many people thought more recent types of memory chips were resistant to. In Monday’s paper, the researchers wrote: “Proprietary, undocumented in-DRAM TRR is currently the only mitigation that stands between Rowhammer and attackers exploiting it in various scenarios such as browsers, mobile phones, the cloud, and even over the network. In this paper, we show how deviations from known uniform Rowhammer access patterns allow attackers to flip bits on all 40 recently-acquired DDR4 DIMMs, 2.6x more than the state of the art. The effectiveness of these new non-uniform patterns in bypassing TRR highlights the need for a more principled approach to address Rowhammer.” While PCs, laptops, and mobile phones are most affected by the new findings, the report notes that cloud services like AWS and Azure “remain largely safe from Rowhammer because they use higher-end chips that include a defense known as ECC, short for Error Correcting Code.”

“Concluding, our work confirms that the DRAM vendors’ claims about Rowhammer protections are false and lure you into a false sense of security,” the researchers wrote. “All currently deployed mitigations are insufficient to fully protect against Rowhammer. Our novel patterns show that attackers can more easily exploit systems than previously assumed.”

Read more of this story at Slashdot.

Hydrogen and Hybrids: Toyota CEO Defends Combustion Engines, Saying ‘The Enemy Is Carbon’

This weekend Toyota’s president drove a specially-equipped Corolla powered by an in-house hydrogen engine, reports Bloomberg. “Along with Mazda Motor Corp., Toyota showcased vehicles running on carbon-neutral propellants in a three-hour road race this weekend in Okayama.”

Toyota’s hydrogen-powered car underscores the automaker’s belief that a wide variety of vehicle types — including hybrids and hydrogen-powered cars, in addition to electric vehicles — will play a role in decarbonizing its fleet over the coming decades. That puts the company in contrast to others, such as General Motors Co., Jaguar Land Rover and Volvo Car AB, which say they’ll sell only EVs two decades from now. “The enemy is carbon, not internal combustion engines,” Toyoda said at a briefing Saturday. “We need diverse solutions, that’s the path toward challenging carbon neutrality.”

Toyota says that that different emissions-reducing car technologies are needed for different regions of the world. EVs are a good option for places like Europe, where batteries can be charged with electricity derived largely from renewable sources, the automaker says. Other options, such as hydrogen or hybrids, may be a better fit in other regions.

The technology is separate from the company’s other big bet on hydrogen — hydrogen fuel cells such as those that power the Mirai passenger car. While fuel cells use the chemical reaction between hydrogen and oxygen to generate electricity, which in turn runs a motor, the hydrogen engine burns the element just like gasoline. Traditional engines only need to be tweaked in minor ways, such as changing out the fuel supply and injection systems, to make them capable of running on hydrogen, Toyota Chief Engineer Naoyuki Sakamoto said in a briefing last month. That also makes the technology a way to save some of the hundreds of thousands of jobs making parts related to combustion engines that are predicted to disappear in Japan if the automotive sector makes a full shift to EVs, according to Toyoda.

Read more of this story at Slashdot.

Will Self-Driving Cars Be Able to Handle… Bears?

A wild bear broke into a parked car looking for food. This set AI pundit Lance Eliot a-thinking…

The AI driving system of a self-driving car is always intact. A parked self-driving car is immediately able to become a moving car…. If the self-driving car is making use of its object detection system, even though the autonomous vehicle is parked, the AI driving capability would be alerted at [a hypothetical] pending crash that is about to occur… Depending upon what the AI developers anticipated, the AI driving system might activate the self-driving car and attempt to quickly drive away from the converging human-driven car.

For most makers of self-driving cars, this is an obscure "edge" case. But Eliot imagines a world where a self-driving car is parked next to a forest…

The human hiker has left the autonomous vehicle and has trekked somewhere deep in the woods. A bear meanders into the parking lot, looking for a free meal. If the AI driving system is using its object detection features, the bear would likely be detected. When the bear decides to wander directly toward the self-driving car, the AI driving system might activate the autonomous vehicle and drive away from the bear.

It is unclear if the bear will somehow divine that the self-driving car is capable of moving on its own accord… After a while, it seems plausible to suggest that bears will be concerned that those free meal containers (on wheels) seem to move away upon the bear approaching. This will possibly discourage some bears and they will steer clear of parked cars. Other bears might turn this into a game. Kind of hide-and-seek, of sorts. Approach a car, it moves away. Fun! Walk over to the car and see which way it goes next. A grand old time in the parking lot, that’s for sure.

And as long as we’re telling shaggy bear stories…
The odds are that self-driving cars will be designed differently on the interior than are conventional human-driven cars. For example, there is no need for a steering wheel and nor any need for the pedals. Those will no longer be included. The interior is opened up to allow for perhaps swiveling seats, possibly reclining seats so that you can sleep on a long journey inside a self-driving car. Given that type of interior, the bear is bound to find things a lot more comfortable inside a self-driving car than a conventional human-driven car. The next thing you know, bears will fall in love with self-driving cars, doing so because it is a quiet, spacious, and secure place to rest and relax. No need to worry about predators getting at the bear while relishing the plush and roomy interior….

A second question is whether the bears might figure out how to communicate with the AI driving system. You know, bears are pretty sharp. Perhaps a truly enterprising bear could convince the AI to take the bear for a cozy ride while inside the self-driving car.

Don’t be especially surprised if you start to see bears riding around in self-driving cars.

And please remember, you heard about it here, first.

Microsoft Is Very Determined That Kids Will Learn To Code Using Minecraft

theodp writes: On Tuesday, Code.org announced that the new activities for kids in this year’s Hour Of Code will include yet another Minecraft-themed tutorial from Code.org Diamond Supporter Microsoft, making it seven years in a row that the best-selling videogame of all time has ‘headlined’ the Hour of Code during the holiday buying season. Going into the Hour of Code in 2018, Microsoft boasted that 100+ million Minecraft Hour of Code tutorials had already been logged by students.

In this year’s Hour of Code: TimeCraft tutorial, kids will “learn basic coding concepts to correct mysterious mishaps throughout history!” An accompanying one-size-fits-all lesson plan for ages 6-18 instructs students to: “Experience a choose-your-own-adventure game, exploring key moments in human achievement. Using your coding superpowers, save the future by solving mysterious mishaps in time.” Among other things, the coding challenges have K-12 students travel back in time to save Jazz from a kazoo future, prevent the Great Pyramids from being built as cubes, save the Great Wall of China from destruction by pandas, and wipe the frown off of the Mona Lisa. New this year, Microsoft notes, is that educators can sign up to have a Microsoft Education Expert lead their classroom through an Hour of Code lesson with Minecraft, thanks to the magic of Microsoft Teams Live Events.

Read more of this story at Slashdot.