Connect with us

Tech

Twitter bans Trump

Published

on

Twitter bans Trump


That means that losing access to the mainstream platforms will reduce his audience and dilute the reach of his statements, as the deplatforming of far-right figures like Alex Jones and Milo Yiannopoulos shows. Yiannopoulos, who was banned in 2016 for his repeated racist abuse of actress Leslie Jones, complained about the effect that deplatforming had on his income.  

“Part of it is because people just don’t remember to go to other websites,” says Joan Donovan, the research director at Harvard’s Shorenstein Center on Media, Politics, and Public Policy. Donovan, a regular MIT Technology Review contributor, points out that the mainstream platforms have built in “bells and whistles” designed to minimize friction and make engaging with content as easy as possible. If Trump were limited to a niche service with limited design and features, such as Parler, she says, it would create an additional barrier to sharing his content. 

Communicating through proxies—with smaller followings

Even during @realdonaldtrump’s day-long absence from Twitter, Trump was not entirely silent on the platform. On Thursday, while the president was still unable to post from his personal account, White House social media director Dan Scavino tweeted a statement from the president that conceded the election—but did not concede his claim that the election was stolen. It was picked up by the media, but with 40,000 retweets and 100,000 likes, it fell far short of the hundreds of thousands that typically engage with each of Trump’s own missives.

As a result, it is “casual supporters” that Trump is most likely to lose if he is permanently banned, says Brooking; they “will hear from him less frequently,” which could mean that that “in time, they may become less wedded to the conspiracy theories and falsehoods that he has made a habit of spreading.”

Of course, it depends on whom he’s speaking through. Much of his disinformation around voter fraud, for example, came from a wider “network of content creation,” says Donovan; that is, individuals close to the president who each have large followings themselves, including Rudy Giuliani, Sidney Powell, and Lin Wood, among others. “These are the accounts that I’m most worried about, because these are the people that are incentivized … because they’re making money off of this,” she says.

 A Trump “digital media empire” could also be blocked

One route around losing his perch on major social media sites could be for Trump to spin up his own systems to talk directly to supporters. The campaign app for his failed campaign for reelection, for example, had its own news and notification system, which often shared questionable or disproven stories that emphasized the president’s talking points.



Tech

Covid clusters among the vaccinated are killing our back-to-normal dreams

Published

on

Covid clusters among the vaccinated are killing our back-to-normal dreams


They were gold miners in French Guiana, revelers in Cape Cod, and Indian health-care workers. Even though they inhabit worlds apart, they ended up having two things in common. All were vaccinated against covid-19. And they all became part of infection clusters.

In recent weeks, cases like these are proving that covid-19 transmission chains and superspreading events can occur even in groups where nearly everyone is vaccinated, setting off alarms among health officials and torpedoing hopes of a quick return to business as usual in the US.

In May 2021, the CDC had told vaccinated Americans they could safety go unmasked, but on Tuesday the agency reversed course, saying vaccinated people should wear masks in indoor public settings.

The reason was what investigators learned from an outbreak in Provincetown, Massachusetts, a seaside town on Cape Cod, which in early July hosted a rowdy parade and crowded weeks of pool parties. Since then, health investigators say, there have been more than 800 cases of covid-19 linked to those events, 74% of which are in people who were vaccinated.

The Provincetown outbreak was caused by the so-called delta variant, which now accounts for most cases in the US.  In a statement released today, Rochelle Walensky, head of the CDC, said the “pivotal discovery” was that vaccinated people infected with delta in Provincetown appear to have just as much virus in their systems as those who are unvaccinated.

“High viral loads suggest an increased risk of transmission and raised concern that, unlike with other variants, vaccinated people infected with delta can transmit the virus,” she said.

The recommendation suggests a rapid return to a layered approach of countermeasures, including masks and social distancing, which could also complicate school reopenings starting next month in the US.

Infection at a gold mine

Investigations around the world have been building evidence of outbreaks among the vaccinated for weeks. For instance, a scientific team in Paris and French Guiana recently described how covid-19 tore through a South American gold mine in May, even though nearly all the miners had received Pfizer’s vaccine.

Despite being inoculated, 60% became infected by a variant called gamma. That surprised the scientists so much that they checked to see if the vaccines had been damaged in shipping, but they weren’t.

Continue Reading

Tech

Boeing’s second Starliner mission to the ISS is a make-or-break moment

Published

on

Boeing’s second Starliner mission to the ISS is a make-or-break moment


Now, Boeing is going for a high-stakes redo of that mission. On August 3, Orbital Flight Test 2, or OFT-2, will send Starliner to the ISS again. The company cannot afford another failure.

“There is a lot of credibility at stake here,” says Greg Autry, a space policy expert at Arizona State University. “Nothing is more visible than space systems that fly humans.”

The afternoon of July 30 was a stark reminder of that visibility. After Russia’s new 23-ton multipurpose Nauka module docked with the ISS, it began firing its thrusters unexpectedly and without command, shifting the ISS out of its proper and normal position in orbit. NASA and Russia fixed the problem and had things stabilized in under an hour, but we still don’t know what happened, and it’s unnerving to think what could have happened if conditions had been worse. The whole incident is still under investigation and has forced NASA to postpone the Starliner launch from July 31 to August 3. 

It’s precisely this kind of near-disaster Boeing wants to avoid, for OFT-2 and any future mission with people onboard.

How Starliner got here

The shutdown of the space shuttle program in 2011 gave NASA a chance to rethink its approach. Instead of building a new spacecraft designed for travel to low Earth orbit, the agency elected to open up opportunities to the private sector as part of a new Commercial Crew Program. It awarded contracts to Boeing and SpaceX to build their own crewed vehicles: Starliner and Crew Dragon, respectively. NASA would buy flights on these vehicles and focus its own efforts on building new technologies for missions to the moon, Mars, and elsewhere. 

Both companies hit development delays, and for nine years NASA’s only way of getting to space was by handing over millions of dollars to Russia for seats on Soyuz missions. SpaceX finally sent astronauts to space in May 2020 (followed by two more crewed missions since), but Boeing is still lagging behind. Its December 2019 flight was supposed to prove that all its systems worked, and that it was capable of docking with the ISS and returning to Earth safely. But a glitch with its internal clock caused it to execute a critical burn prematurely, making it impossible to dock with the ISS. 

A subsequent investigation revealed that a second glitch would have caused Starliner to fire its thrusters at the wrong time when making its descent back to Earth, which could have destroyed the spacecraft. That glitch was fixed mere hours before Starliner was set to come back home. Software issues aren’t unexpected in spacecraft development, but they’re things Boeing could have resolved ahead of time with better quality control or better oversight from NASA.

Boeing has had 21 months to fix these problems. NASA never demanded another Starliner flight test; Boeing elected to redo it and foot the $410 million bill on its own.

“I fully expect the test to go perfectly,” says Autry. “These problems involved software systems, and those should be easily resolvable.”

What’s at stake

If things go wrong, the repercussions will depend on what those things are. Should the spacecraft experience another set of software problems, there’ll likely be hell to pay, and it’s very hard to see how Boeing’s relationship with NASA could recover. A catastrophic failure for other reasons would also be bad, but space is volatile, and even tiny problems that are hard to anticipate and control for can lead to explosive outcomes. That may be more forgivable.

If the new test doesn’t succeed, NASA will still work with Boeing, but a re-flight “might be a couple years off,” says Roger Handberg, a space policy expert at the University of Central Florida. “NASA would likely go back to SpaceX for more flights, further disadvantaging Boeing.”

Boeing needs OFT-2 to go well for reasons beyond just fulfilling its contract with NASA. Neither SpaceX nor Boeing built its new vehicles to carry out ISS missions—they each had larger ambitions. “There is real demand [for access to space] from high-net-worth individuals, demonstrated since the early 2000s, when several flew on the Russian Soyuz,” says Autry. “There is also a very strong business in flying the sovereign astronaut corps of many countries that are not ready to build their own vehicles.”

SpaceX will prove to be very stiff competition. It has private missions—its own and through Axiom Space—already slated for the next few years. More are sure to come, especially since Axiom, Sierra Nevada, and other companies plan to build private space stations for paying visitors. 

Boeing’s biggest problem is cost. NASA is paying the company $90 million per seat to fly astronauts to the ISS, versus $55 million per seat to SpaceX. “NASA can afford them because after the shuttle problems the agency did not want to become dependent upon a single flight system—if that breaks, everything stops,” says Handberg. But private citizens and other countries are likely to plump for the cheaper—and more experienced—option.

Boeing could definitely use some good PR these days. It is building the main booster for the $20-billion-and-counting Space Launch System, set to be the most powerful rocket in the world. But high costs and massive delays have turned it into a lightning rod for criticism. Meanwhile, alternatives like SpaceX’s Falcon Heavy and Super Heavy, Blue Origin’s New Glenn, and ULA’s Vulcan Centaur have emerged or are set to debut in the next few years. In 2019, NASA’s inspector general looked at potential fraud in Boeing contracts worth up $661 million. And the company is one of the main characters at the center of a criminal probe involving a previous bid for a lunar lander contract. 

If there was ever a time Boeing wanted to remind people what it’s capable of and what it can do for the US space program, it’s next week.

“Another failure would put Boeing so far behind SpaceX that they might have to consider major changes in their approach,” says Handberg. “For Boeing, this is the show.”

Continue Reading

Tech

Hundreds of AI tools have been built to catch covid. None of them helped.

Published

on

Hundreds of AI tools have been built to catch covid. None of them helped.


It also muddies the origin of certain data sets. This can mean that researchers miss important features that skew the training of their models. Many unwittingly used a data set that contained chest scans of children who did not have covid as their examples of what non-covid cases looked like. But as a result, the AIs learned to identify kids, not covid.

Driggs’s group trained its own model using a data set that contained a mix of scans taken when patients were lying down and standing up. Because patients scanned while lying down were more likely to be seriously ill, the AI learned wrongly to predict serious covid risk from a person’s position.

In yet other cases, some AIs were found to be picking up on the text font that certain hospitals used to label the scans. As a result, fonts from hospitals with more serious caseloads became predictors of covid risk.

Errors like these seem obvious in hindsight. They can also be fixed by adjusting the models, if researchers are aware of them. It is possible to acknowledge the shortcomings and release a less accurate, but less misleading model. But many tools were developed either by AI researchers who lacked the medical expertise to spot flaws in the data or by medical researchers who lacked the mathematical skills to compensate for those flaws.

A more subtle problem Driggs highlights is incorporation bias, or bias introduced at the point a data set is labeled. For example, many medical scans were labeled according to whether the radiologists who created them said they showed covid. But that embeds, or incorporates, any biases of that particular doctor into the ground truth of a data set. It would be much better to label a medical scan with the result of a PCR test rather than one doctor’s opinion, says Driggs. But there isn’t always time for statistical niceties in busy hospitals.

That hasn’t stopped some of these tools from being rushed into clinical practice. Wynants says it isn’t clear which ones are being used or how. Hospitals will sometimes say that they are using a tool only for research purposes, which makes it hard to assess how much doctors are relying on them. “There’s a lot of secrecy,” she says.

Wynants asked one company that was marketing deep-learning algorithms to share information about its approach but did not hear back. She later found several published models from researchers tied to this company, all of them with a high risk of bias. “We don’t actually know what the company implemented,” she says.

According to Wynants, some hospitals are even signing nondisclosure agreements with medical AI vendors. When she asked doctors what algorithms or software they were using, they sometimes told her they weren’t allowed to say.

How to fix it

What’s the fix? Better data would help, but in times of crisis that’s a big ask. It’s more important to make the most of the data sets we have. The simplest move would be for AI teams to collaborate more with clinicians, says Driggs. Researchers also need to share their models and disclose how they were trained so that others can test them and build on them. “Those are two things we could do today,” he says. “And they would solve maybe 50% of the issues that we identified.”

Getting hold of data would also be easier if formats were standardized, says Bilal Mateen, a doctor who leads research into clinical technology at the Wellcome Trust, a global health research charity based in London. 

Continue Reading

Copyright © 2020 Diliput News.