Connect with us

Science

Apple adding its own co-processors to three Macs in 2018, Bloomberg reports

Published

on


Apple’s own T2 co-processor inside the iMac Pro desktop.

A new Bloomberg article is reporting that Apple plans to expand the number of Macs using its own co-processor chips this year, a move that could foretell a future where Intel is no longer inside Apple’s computers.

According to the report, Apple is looking to add one of its custom co-processing chips to a pair of MacBook laptops and a Mac desktop to be released later this year. They would join the existing MacBook Pro with Touch Bar and iMac Pro as being Mac systems with Apple-made co-processors. Of course, iPhones and iPads have used the company’s own chips for several years, and the Apple Watch has since its launch in 2015.

For now, Intel remains the provider of the main CPU for Macs, though Bloomberg claims that the recent security woes that have dogged that company’s chips may have provided Apple more ammunition to chart its own course for future Mac processors. Another report today claims Apple is slowing its roll-out of new iOS features in 2018 as it deals with reliability issues, so security — something the company has long touted compared to systems without its “walled garden” approach — is receiving greater attention in Cupertino.

Apple avoids some of the cost issues that other companies have had in the past building their own chips by outsourcing the manufacturing. Designing chips in-house also allows the company to tailor them to new features it’s developing. Abandoning its on-again off-again relationship with Intel would starve the chip maker of its fifth-largest client, according to Bloomberg.

Despite the increasing resources Apple is devoting to chip research and creation, it’s only built computer chips that complement Intel’s primary processor — the T1 co-processor handles the Touch Bar, while the T2 edition offloads some security and power management duties on the iMac Pro. It remains to be seen what a potential T3 co-processor will control, and there’s quite a leap from a chip working on a few tasks to ones that have to power the entire system, potentially including graphics processing to boot.

If Apple decides in the future to produce its own CPUs for Macs, it could drive a great debate among its fanboys. Would Macs (and their users) be better served by Intel, whose high-performance chips have been undermined by security holes, or by Apple, whose chips would be untested in laptops and desktops but could maximize the potential of the Mac OS? It seems unlikely we’ll find out in 2018, but another year of behind-the-scenes work by Apple could make for a very interesting 2019.

Source link



Source link

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Science

The US is doing well on emissions, but not “halfway to zero”

Published

on

On Monday, the US Department of Energy’s Lawrence Berkeley Lab released a report entitled “Halfway to Zero,” referring to a goal of a zero-emission US electric grid. The report’s headline claim is a bit bogus, in that we’ve not cut our emissions in half relative to any point in history. Instead, they’re down to half of where they were projected to be in a report issued back in 2005.

Still, despite this sleight-of-hand, the report makes for interesting reading in that it shows how rapidly the energy market has changed and where the trends that are driving those changes might lead us in another 15 years. So, while we’re not really in a place to be patting ourselves on the back for everything we’ve accomplished, the report does provide reasons for optimism.

Persistently wrong projections

The foundation for the new work is one of the many editions of the US Energy Information Agency’s (EIA) Annual Energy Outlook, specifically the one from 2005. These publications take a look at the state of the US energy markets in the most recent year for which there is data and tries to project how those markets will evolve over the coming years. In 2005, that meant that projections went as far as 2025.

You can get a sense of where things were expected to go simply by looking at the cover of the report, which is graced by a picture of an oil well. To an extent, conservative projections are baked into the structure of the report. It assumes that the only policies that influence energy markets are the ones currently on the books. So, even though Congress has extended the expiration date on tax credits for renewable power multiple times, the EIA always makes projections that have them expiring as scheduled at the time.

The EIA also tends to assume slow and steady technological progress, rather than the sorts of changes that have led solar prices to drop by a factor of five in less than a decade.

So, the EIA projections will be wrong, and everyone knows they’ll be wrong and has a good sense of why. The new report acknowledges that very early on and just moves on to the comparisons. And the comparisons it makes are informative in that they show what factors are the leading causes of the mismatch between the projections and the reality that has since taken place.

In this case, however, there’s an added wrinkle: the COVID-19 pandemic, which has lowered energy use in the US considerably. It’s likely that at least some of the factors that drove the changes in energy use, like increased work from home, will outlast the pandemic, but it’s difficult to know how different things will be. While the report also performs an analysis using 2019 data to avoid this complication, we’ll largely focus on the 2020 results.

Demand is less demanding

One of the biggest factors that has changed what’s happened on the grid is a drop in demand compared to expectations. In its 2005 projections, the EIA had expected total demand for electricity to rise by roughly 25 percent. Instead, it stayed essentially flat. None of the main sectors—residential, commercial, or industrial—rose significantly, due to factors like increased efficiency and a shift away from heavy industry. (Remember, things like LED bulbs were expensive rarities 15 years ago.)

This had some significant effects. For example, the EIA projected a large increase in natural gas use by 2020 and got the magnitude nearly right. But, in its projections, this expanded generation was expected to feed the rising demand; since that never materialized, gas displaced coal instead.

Rather than rising as projected, demand has largely remained flat.
Enlarge / Rather than rising as projected, demand has largely remained flat.

Coal also took a big hit from the rise in renewables, with wind and solar producing 13 times more than they were expected to back in 2005. Combined with hydropower and other renewable sources, that’s boosted the total share of renewable generation to 79 percent higher than the EIA had projected. There are plenty of reasons for this, including the fact that tax breaks for renewables were extended several times. In the absence of a coordinated federal response to climate change, many states have crafted renewable power mandates. And, finally, there’s that tremendous drop in price noted above, which has made solar and wind the cheapest sources of power in many areas of the country.

Combined, this has produced a serious drop in carbon emissions: 40 percent since 2005, or about a third of the way toward a zero-emissions grid. Compared to the 2005 projections for this year, which had pencilled in continued heavy use of coal and rising demand, emissions are down by 52 percent—the “halfway to zero” of the report’s title.

Counting costs

In 2005 (and for a number of years afterwards), people frequently announced that reducing the carbon emissions of the grid would come at a catastrophic cost. That’s turned out to be very badly wrong. Retail electricity prices (in 2005 constant dollars) went from 10.6 cents/kiloWatt-hour all the way up to… 10.7 cents/kW-hr. The projections had them falling slightly, so we’re in worse shape than expected in this sense, but it’s anything but a crippling rise in price.

In fact, because there are more customers for a similar amount of power, use per customer has gone down in the last 15 years. That means that, individually, customers’ bills are a bit smaller, even if they’re paying more.

And that’s been accompanied by a huge drop in associated costs. Climate damages, estimated using the social cost of carbon, were cut by more than half, from a projected $229 billion down to only $110 billion. But the huge plunge in the use of coal, along with added pollution controls, cut down on numerous additional costs. Health costs from the pollution, expected to total over a half-trillion dollars, instead came in at $34 billion—an astonishing drop. The projected 38,000 annual premature deaths from pollution instead came in at 3,100.

Any way you look at it, carbon emissions are down.
Enlarge / Any way you look at it, carbon emissions are down.

If these costs are counted along with the electrical bills, the total costs of generating power were actually down by 44 percent from 2005 and less than half of what was projected for 2020.

Despite the fact that the cost of the energy is largely the same in constant dollars, the energy industry is employing a lot more people. Despite the boost in production, the number of jobs in natural gas production has stayed largely the same, while those employed producing coal has dropped considerably. But renewable energy production is labor-intensive, which has led to a dramatic jump in jobs and brought the total number of people employed in producing electricity up 29 percent compared to projections.

Where to from here?

While we may not be halfway to a net zero emissions grid, we’re a lot closer than might be expected. Are we on a trajectory that makes zero emissions possible within the sorts of time frames we’d need in order to hit climate goals? The report acknowledges that decarbonizing the last 10 percent of the grid using intermittent power sources is going to be an extremely complex challenge, and likely involve some combination of greater grid integration, demand management, storage, and carbon capture. But we have a lot of decarbonization to do before we even get to the 10 percent mark, and it’s here that we can largely focus on trends in existing technology.

The report notes that wind, solar, and batteries have all gotten cheaper much faster than anyone expected, making deep decarbonization an economic possibility. But there’s still the challenge of scaling production and installation of renewables and storage fast enough.

To get to a situation where 90 percent of the electricity generated in the US is free of carbon emissions, it’s estimated we’ll need to install about 1.1 terawatts of wind and solar capacity by 2035. Right now, there’s been applications to connect roughly half that amount by the end of 2025, a pace of over 100 GW/year, well above what’s needed. But historically, many of these projects don’t end up getting built; as a contrast, only 34 GW was installed in 2020, which is only a third of the expected pace.

We’d need to double last year’s total starting this year and maintain that pace for a decade to hit the 90 percent emissions-free goal. Plus we have to build enough grid infrastructure to connect it all. That’s not going to be easy. For a sense of how difficult it will be, compare the 70 GW/year figure to the big plan to expand offshore wind production in the US, which is only targeting three GW/year for the rest of this decade.

Achieving the sort of scale we need is going to be the real challenge if we want to hit our climate goals. And so far, even with a renewable-friendly administration in the White House, there’s been no indication of how we might do it.

Continue Reading

Science

The very common vaccine ingredient at the center of J&J, AstraZeneca drama

Published

on

Enlarge / Adenoviruses seen via transmission electron microscopy.

Out of an abundance of caution, US officials on Tuesday recommended pausing use of Johnson & Johnson’s COVID-19 vaccine. Officials linked the vaccine to six peculiar illnesses in which people developed life-threatening blood clots in combination with low levels of blood platelets, the cell fragments in blood that form clots. One person died from their condition and another is in critical condition.

It’s unclear if the vaccine caused the illnesses. Even if it did, the illnesses would represent an exceedingly rare side effect. The six cases occurred among more than 6.8 million people in the US who received the Johnson & Johnson vaccine. That would make it a side effect seen in fewer than one in a million. The risk of hospitalization and death from COVID-19, which the vaccine protects against, easily exceeds those odds. Without question, the benefits of the vaccine outweigh the potential risks.

Still, with robust supplies of vaccine from Moderna and Pfizer-BioNtech—neither of which have been linked to these unusual cases—US officials took the cautious route of pausing Johnson & Johnson’s vaccine while they investigate the cases further and inform clinicians about how to spot and treat any others that may arise. This latter point is critical because if doctors try to use standard blood clot treatments in these vaccine-linked cases, the outcomes can be fatal.

Of course, the other critical aspect of this situation is that officials have seen these unusual cases before—linked to a similar COVID-19 vaccine developed by AstraZeneca and researchers at the University of Oxford. The AstraZeneca vaccine is not yet authorized for use in the US, but it has been authorized in many other countries, including those in the European Union. In recent weeks, regulators in the EU and the UK have investigated dozens of eerily similar cases, involving dangerous blood clots coupled with low platelets. Some estimates have pegged the reported case rate of one in 100,000 people vaccinated.

Trying now to connect all the dots and find answers, experts are eyeing the most obvious connection: both vaccines use an adenovirus vector, a viral delivery system used regularly in vaccine development.

At the moment, the adenovirus vector offers “the most straightforward explanation” for the possible side effects, says viral immunologist Hildegund Ertl, who develops adenovirus-based vaccines at the Wistar Institute in Philadelphia. Yet, the link to blood clots “took all of us by surprise,” she tells Ars. The situation has raised a slew of questions—as well as some doubts.

Vexing virus

Adenoviruses are a large family of very common viruses that cause a range of infections in humans, from mild colds and flu-like illnesses to pink eye, pneumonia, and gastroenteritis. Beyond humans, they can infect a range of animals, including pigs, cows, and chimpanzees. Researchers have been working with them for decades. The Johnson & Johnson vaccine uses the adenovirus (Ad26), which was first identified in 1961 from anal swabs of children in Washington, DC. The AstraZeneca vaccine is based on an adenovirus that circulates in chimpanzees (ChAdOx1).

Over the years, researchers have considered adenoviruses useful delivery systems for vaccines and gene therapies. For starters, they’re easy to brew up in big batches in laboratory conditions. When engineered for vaccines, they can provoke potent immune responses in people against germs we want to fight. And they appeared relatively safe in humans, particularly since they’re often modified so they can’t replicate in our cells.

But adenoviruses have had a troubled past. Researchers all but abandoned their use in gene therapies in 1999 following the tragic death of 18-year-old Jesse Gelsinger. A team of researchers at the University of Pennsylvania had hoped to cure the teenager’s rare metabolic liver disease by correcting an underlying genetic mutation with new code—delivered in trillions of adenovirus vectors. The researchers used human adenovirus 5 (Ad5), which typically causes only a mild cold. In early tests, the therapy triggered only mild side-effects and flu-like symptoms in animals and a human patient, the New York Times reported at the time. But in Gelsinger, the massive dose of virus vectors triggered a fatal immune response.

Researchers carried on with adenoviruses for vaccine development, where potent immune responses can be a plus instead of a peril. Programmed to be vaccine vectors, adenoviruses deliver key snippets of genetic code from dangerous viruses, bacteria, or parasites directly to human cells. From there, our cells translate the genetic code into protein, recognize it as foreign, and use it to train our immune systems to seek and destroy anything carrying the same protein. In the case of COVID-19, adenovirus-based vaccines carry the genetic code for the SARS-CoV-2 spike protein, which is the thorny protein that juts from the virus’s particle. The spike protein is what SARS-CoV-2 uses to enter human cells, and it’s a key target for potent antibodies and other immune responses.

Vials of the AstraZeneca COVID-19 vaccine are seen during the opening of a vaccination center in Cyprus on March 22, 2021.
Enlarge / Vials of the AstraZeneca COVID-19 vaccine are seen during the opening of a vaccination center in Cyprus on March 22, 2021.

Shaky shots

Adenovirus-based vaccines have held a lot of promise over the years, but they have had notable stumbles, too. Nearly a decade after Gelsinger’s death, researchers halted a major trial of an Ad5-based HIV vaccine after data indicated that the vaccine increased the risk of becoming infected with HIV in people who had preexisting immune responses to Ad5. With the high-profile failure, many vaccine developers moved away from Ad5 to other adenoviruses—ones that people tend to have less preexisting immunity against, like chimpanzee adenoviruses.

Though researchers have been developing adenovirus-based vaccines against a slew of diseases—malaria, HIV, Zika, RSV (respiratory syncytial virus) and more—few have made it across the finish line and into use. Among the most successful is an Ad26-based Ebola vaccine made by Johnson & Johnson, which gained regulatory approval in Europe last year. The approval bolstered hopes for the company’s COVID-19 vaccine, which uses the same Ad26-based platform.

Early on in the pandemic, the adenovirus-based vaccines were often seen as front-runners, particularly AstraZeneca’s. Despite the checkered past of adenovirus vectors, the vaccine design was seen as a more established technology than the mRNA-based vaccines, which were completely unproven until the extraordinary success of COVID-19 vaccines from Moderna and Pfizer-BioNTech. Adenoviruses also have logistical advantages. They’re relatively cheap, easy to make, and easy to distribute. For instance, unlike the mRNA vaccines which require ultra-cold storage conditions, AstraZeneca’s vaccine can handle normal refrigerator temperatures. Many experts and the World Health Organization have considered AstraZeneca’s vaccine as the world’s go-to vaccine—a cheap, accessible vaccine that could be used in a range of countries and settings.

But as the mRNA vaccines sprinted ahead in the pandemic, AstraZeneca seemed to lurch from problem to problem. The vaccine’s troubles hit a critical point last month when more than a dozen countries temporarily suspended use of AstraZeneca’s vaccine amid concerns that they were causing extremely rare blood clots. On April 7, an investigation by the EU’s European Medicines Agency concluded that there was a strong association between the vaccine and peculiar illnesses involving both blood clots and low platelets. The agency determined that they should be listed as “very rare side effects” of the vaccine, but it still urged countries to continue using the vaccine.

“The reported combination of blood clots and low blood platelets is very rare,” the agency noted. “The overall benefits of the vaccine in preventing COVID-19 outweigh the risks of side effects.”

Continue Reading

Science

Tesla tells customers they’ll have to pay more for solar roof

Published

on

Tesla has jacked up the price of its solar roof, which integrates solar panels directly into roof tiles, Electrek reports. A 12.3 kW system that Electrek priced at $54,966 last summer now costs more than $70,000, according to Tesla’s online calculator.

Customers report that Tesla is not only raising prices for future solar roof installations—it’s demanding more money from some existing customers whose panels haven’t been installed yet.

“Tesla has reneged on its solar roof contracts and has raised the prices for people (like me) who have already signed contract with them,” one Ars reader told us. “For example, they are asking for an additional $21,000 for my project from about $66,000 to $87,000.”

Another customer shared a similar story with The Verge:

One solar roof customer told The Verge he signed a contract in February to install the solar roof for $35,000, with an additional $30,000 for the batteries. Later, he received a “terse email” from Tesla stating that in several days he would receive a new contract with higher prices. He was told he now owed $75,000 for the solar roof—a 114 percent increase—and $35,000 for the batteries.

Part of the price increase may reflect Tesla’s realization that not all roofs are equally easy to cover with solar panels. Tesla’s online cost estimator now asks customers about their “roof complexity,” which is “determined by the pitch, number of joints, chimneys, and other features on your roof.” Tesla warns that “complex roofs take more time and material to install than simple roofs.”

If Tesla’s online price calculator for solar roof is accurate, its new, higher prices are still fairly competitive with conventional solar panels—at least here in the District of Columbia. In 2019, I paid $35,538 for an 8.5 kW system, spending $12,000 for a new roof at the same time. Today, Tesla estimates that a 8.5 kW system would cost $36,783 for a “simple” roof. That compares favorably to the $47,538 I paid for solar panels and a new roof 18 months ago. Of course, solar panel prices have continued to drop in the last 18 months, so conventional solar panels would likely cost less if I ordered them today.

Tesla’s solar roof apparently can’t achieve the same energy density as conventional solar panels. I installed 8.5 kW worth of conventional solar panels on my roof, which is around 1,000 square feet. In contrast, Tesla’s calculator indicates that it takes more than 1,400 square feet of roof to fit 8.5 kW worth of solar roof tiles.

There’s also no guarantee that Tesla won’t change its estimate once it sends someone to actually look at your house. The cost calculated by Tesla’s website is only an initial estimate.

Solar roof has a troubled history

Tesla’s solar roof dates back to 2016, the year Tesla bought SolarCity, a solar panel installation company founded by two of Tesla CEO Elon Musk’s cousins. Prior to the merger, SolarCity installed conventional solar panels that sit above a customer’s roof (Tesla continues to install conventional solar panels today). But in 2016, weeks before the SolarCity deal was due to close, Tesla announced the solar roof project, which features roofing tiles that double as solar panels.

Tesla started accepting preorders for solar roof in May 2017. In August 2018, Musk told investors that Tesla had “several hundred solar roofs” installed. But the company then quickly walked that statement back, clarifying that Musk was including people who were scheduled for roof installations but hadn’t gotten them yet. The same month, Reuters reported that Tesla was struggling to manufacture roof tiles at its solar panel factory in Buffalo, New York.

In 2019, Tesla announced it was slashing prices for the solar roof in an effort to boost sluggish demand for the product. In 2020, Tesla canceled the orders of some customers because they were out of Tesla’s service area. At the same time, Tesla said it had “roughly tripled” solar roof installations between the first and second quarters of 2020.

Tesla’s Q3 2020 earnings report bragged that Tesla had reduced the time required to install a solar roof to as little as 1.5 days. “We continue to onboard hundreds of electricians and roofers to grow this business,” Tesla wrote. Tesla’s Q4 2020 earnings report stated that the company “made great progress growing our Solar Roof deployments,” but it didn’t provide any hard numbers.

All of which is to say that, after years of delays and setbacks, Tesla may be finally selling the solar roof at significant scale. And the company has apparently discovered that the costs of installing a solar roof are higher than previously anticipated for many customers.

Continue Reading

Trending