Topps Blog

Tuesday, 27 December 2022

Scientists Use Powerful New Climate Model to Recreate Iconic ‘Blue Marble’ Photograph

This month marks 50 years since humanity last ventured beyond low-Earth orbit. While NASA is currently working to send people back to the moon, Apollo 17 in 1972 is the most recent crewed mission to leave Earth behind. The mission is notable for another reason: the famous “Blue Marble” photograph of Earth fully illuminated by the sun. This stunning image supercharged the nascent environmental movement in the early 1970s and has been reproduced frequently over the decades. Now, it’s being reproduced in a new way — scientists from the Max Planck Institute for Meteorology attempted to recreate the Blue Marble shot from scratch using a new climate model. The results, as you can see above, are impressive. The simulation on the right is a dead ringer for the authentic photo on the left.

You could argue that the foundation of the Max Planck Institute for Meteorology (MPI-M) was part of the environmental movement spurred on by the Blue Marble. MPI-M was set up under the direction of Prof. Klaus Hasselmann in 1975, who would go on to win the 2021 Nobel Prize in Physics for MPI-M’s work on anthropogenic climate change.

Work on the climate at MPI-M is still ongoing. In late 2022, the institute completed work on a model known as ICON that can simulate fully coupled climate systems at kilometer scale. Climate scientists believe that being able to resolve atmospheric processes at this level of detail is essential to understanding how global warming is reshaping the globe. With the 50-year anniversary of the iconic Blue Marble photo, the MPI-M team decided to test the model by recreating the image using weather data from 1972.

MPI-M partnered with Nvidia and the German Climate Computing Center to run the simulation, beginning with a spun-up ocean two days before the famous image was captured. The institute describes it as a two-day forecast arriving fifty years too late. By simulating the physics of surface winds, water currents, and cloud fields, the team ended up with something that very closely matches the real photograph.

The image above was generated with the Nvidia Omniverse platform with internal RTX ray tracing. And it’s not just a photograph. The simulation is a fully formed world, featuring details not recorded by the Apollo crew’s camera. The simulation allowed the team to go beyond the superficial, studying the waves of warm water emanating from the African coast and subsurface eddies that rise upward as the sun heats the ocean. The team expressed pride that ICON was able to so closely mirror the Blue Marble, particularly given the rudimentary weather data from 50 years ago in the southern hemisphere. This work shows that ICON could become a key tool in understanding Earth’s changing climate.

Now read:

Intel May Cancel Meteor Lake Desktop CPUs in Favor of Raptor Lake Refresh

(Credit: Intel) Intel famously stalled out at 14nm for many years, allowing AMD to leapfrog it in the process with technology from TSMC. Now it might be experiencing that issue once again trying to make the jump from 10nm to 7nm.

Put another way, Intel is supposed to be launching its first tile-based CPU next year, aka Meteor Lake. This is the company’s first disaggregated CPU, with different tiles built on differing processes. The CPU tile is supposed to be the result of a node jump, going from Intel 7 (10nm) to Intel 4 (7nm). It’s a watershed product for the company as it transitions away from monolithic and hybrid designs to tiles. However, though Meteor Lake has already been delayed many times, it’s now rumored to be cancelled. At least, the desktop parts are believed to be on the chopping block. Therefore, it might still launch at some point, but only for mobile.

The bad news for Chipzilla comes from Twitter user Raichu, a reliable leaker of confidential information, who says Meteor Lake for desktop is “maybe” cancelled because Intel sees it as uncompetitive. Intel had planned on offering only six performance cores, along with 16 efficiency cores. That was laid out in a previous leak of Intel’s plans. As you recall, its current Raptor Lake flagship has eight performance cores, and 16 efficiency cores, so Meteor Lake would be a step backward. It could be true that Intel doesn’t think it’s competitive, and it could also be true it’s delayed once again. It’s not clear the specific reason for the alleged cancellation. However, it is now planning a Raptor Lake refresh in 2023, as we noted previously.

(Credit: Intel)

The leaked roadmap for 2023 above doesn’t mention Meteor Lake. Instead, Intel will be offering a Raptor Lake refresh on a refined 10nm process. That should allow it to squeeze a few hundred megahertz out of the chips, and improve its efficiency a bit. The good news for Intel fans is it will use the same LGA 1700 socket it’s currently offering. There were rumors that Meteor Lake would come out after Raptor Lake on a new LGA 1851 socket. That would give LGA 1700 only two generations of CPUs. Now it appears it’ll be three: Alder Lake, Raptor Lake, and the Raptor Lake refresh.

Intel’s previous roadmap might have changed just a tad. (Credit: Intel)

As you recall, Raptor Lake only exists because Intel realized over two years ago that Meteor Lake was not hitting its development targets. That was confirmed by Intel, so it’s not hearsay. Fast forward to now in 2022, and nobody would be surprised if it’s still having trouble with it. Almost every article we’ve written about Meteor Lake has concerned delays. There were rumors it was cancelling the 3nm GPU tile from TSMC due to delays. Instead, it would use the company’s existing 5nm node. It was also reported in July that Meteor Lake was delayed to the end of 2023. If you scroll up and look at those roadmaps, you can see Meteor Lake has disappeared from the 2023 timeline.

(Credit: Intel)

Now it seems like Intel feels Raptor Lake will be more competitive than Meteor Lake against AMD’s upcoming offerings. That’s expected to be more Zen 4 CPUs along with several V-Cache SKUs. Also, perhaps its current tile-based plans will be better for efficiency than outright horsepower. That would explain Meteor Lake being a mobile platform instead of for desktop. So Arrow Lake would take up the mantle on the desktop. That’ll happen sometime in 2024 on Intel’s 20A process. Arrow Lake is alleged to resume the 8/16 core count it’s using currently.

In an October earnings call, CEO Pat Gelsinger made the following statement, according to Wccftech: “On Intel 4, we are progressing towards a high-volume manufacturing and will tape out the production stepping at Meteor Lake in Q4. Intel 4 and 3 are our first nodes deploying EUV and will represent a major step forward in terms of transistor performance per watt and density.”

If that’s true, that means Intel is already producing early versions of Meteor Lake right now. We’ll have to wait and see if that’s the case, but Intel doesn’t usually respond to allegations like these.

Now Read:

A Geoengineering Startup Is Releasing Sulfur Into the Atmosphere, Selling ‘Cooling Credits’

(Credit: RUNSTUDIO/Getty Images) The Earth’s climate is warming, and the potential for serious consequences in the future is strong. Are things so dire that we need to start manipulating the atmosphere? One startup thinks so, and it’s moving forward with a plan that scientists have condemned. Make Sunrise has started releasing reflective particles to cool the globe, and it’ll even sell you “cooling credits” as a byproduct. It sounds good on paper, but this early attempt at geoengineering risks causing more problems than it solves.

Make Sunsets uses weather balloons that climb high into the stratosphere, releasing small clouds of sulfur particles. According to the company, one gram of sulfur released into the upper atmosphere counteracts one ton of carbon emissions for a period of one year. Scientists have been studying this form of geoengineering in the abstract for years — it mimics the natural release of sulfur in volcanic eruptions, which can lower temperatures as the particles reflect small amounts of sunlight. But doing it intentionally could have unforeseen consequences.

Multiple experts in the field of geoengineering contacted by MIT Technology Review have heavily criticized the company for moving forward with its plans. Luke Iseman, CEO of Make Sunsets, isn’t surprised. He says that the company is equal parts entrepreneurial and provocation. He also understands the reaction, noting that it makes him look like a Bond villain. And like all good Bond villains, Iseman has a foreign lair from which he carries out his dastardly plan — in this case, Baja California.

Iseman says the company’s first two balloon launches took place in Baja California this past April, which was before the company officially formed. The balloons were loaded with a few grams of sulfur and enough helium to get them high into the atmosphere, but we don’t know how high. The goal was to get the balloons to burst in the stratosphere, but it’s unclear if that happened or what the impact of the particles was. There was no monitoring hardware on the balloon, and Iseman did not seek the approval of any government before releasing the balloon.

While geoengineering has not been undertaken on a large scale, it’s unlikely that a few grams of sulfur in a balloon will have serious negative impacts. But it’s not certain it will do any good, either. Thus, the $10 Make Sunsets charges per “cooling credit” is completely speculative. The danger is that no one stops Make Sunsets from expanding operations, which could attract other players in the burgeoning field of geoengineering. Before we know it, we may have altered the atmosphere in unexpected ways. And maybe that will mean lower global temperatures, but what else could it mean?

Now read:

California Bans Tesla From Calling Its Software Full-Self Driving

(Credit: Tesla) California Gov. Gavin Newsom has signed a new bill into law that effectively bans Tesla from advertising its cars as fully self-driving, Government Technology reports. The legislation stipulates that beginning Jan. 1, 2023, Tesla can no longer use the phrase Full-Self Driving, commonly abbreviated as FSD, to refer to a feature in its cars—something that we’ve always wondered about, as Tesla vehicles have never progressed beyond Level 2 self-driving ability.

Senate Bill 1398, sponsored by Democratic State Sen. Lena Gonzalez, prohibits dealers and automakers from “deceptively naming or marketing” a car as self-driving if it still requires human attention and intervention. The report said that the state DMV already had rules banning this kind of false advertising, but the lack of enforcement led Gonzalez and state legislators to advance the bill so that it’s now state law.

“(This bill) increases consumer safety by requiring dealers and manufacturers that sell new passenger vehicles equipped with a semiautonomous driving assistance feature… to give a clear description of the functions and limitations of those features,” Gonzalez said in a statement. Tesla has countered by saying that it already makes its customers aware of the software’s limitations.

(Credit: Sjoerd van der Wal/Getty Images)

Currently, Tesla sells FSD as an additional over-the-air option for $15,000 or $199 per month. A statement on Tesla’s website regarding the feature reads: “The currently enabled Autopilot, Enhanced Autopilot, and Full Self-Driving features require active driver supervision and do not make the vehicle autonomous. Full autonomy will be dependent on achieving reliability far in excess of human drivers as demonstrated by billions of miles of experience, as well as regulatory approval, which may take longer in some jurisdictions.” The automaker also says on its website that all new vehicles already have the hardware necessary to add FSD in the future.

Tesla CEO Elon Musk said in 2017 that he had planned for a Tesla to drive itself across the country without the need for human assistance before 2018. As PCMag reports, last week the National Highway Traffic Safety Administration (NHTSA) launched two more crash investigations into accidents where Tesla FSD is alleged to have been at least partially responsible.

Earlier this month, a filing with the Federal Communications Commission (FCC) indicated Tesla is working on a new radar system operating in the 76-77GHz range. The filing suggests the automaker is planning to bring back a radar hardware component to at least some of its models, which raises the question of whether existing vehicles without this hardware will ever be good enough to drive themselves without human intervention.

Now read:

Friday, 23 December 2022

Godfather Android Malware Targets 400+ Banks and Crypto Exchanges

After fading away for several months, the newly prevalent Godfather Android malware is back with a vengeance, targeting more than 400 international financial firms. The trojan generates fake login pages to harvest customer login details, and that’s just the start. Godfather also mimics Google’s pre-installed security tools in an attempt to gain full control over devices.

Godfather was discovered by malware analytics firm Group I-B, with the first samples appearing in June 2021. It is believed this malware grew out of another popular bank hacker known as Anubis. Godfather circulated at low levels until June 2022, when it vanished. It appears the operators were simply preparing a new version. Godfather was back with a vengeance in September of this year, targeting a whopping 400 financial companies: 215 international banks, 94 cryptocurrency wallets, and 110 crypto exchanges.

When installed on a device, Godfather will generate fake login pages, which it can use to get usernames and passwords. Many banks and crypto firms have additional login requirements, and that’s where Godfather’s other mechanisms come in handy. After installation, the malware masquerades as a Google Play Protect alert. Thinking this is a legitimate popup from Android’s default security suite, some users will grant the malware accessibility control. At that point, Godfather can record the screen, read SMS, fire off fake notifications, make calls, and more — everything you need to compromise a bank account or crypto vault.

Godfather’s fake Play Protect popup.

The malware appears to be spreading via decoy apps in the Play Store. Group I-B has not determined who created and profits from Godfather, but it heavily suspects that they are Russian speakers. There’s a kill switch in the malware that checks the OS language setting. If it finds the default language is one of those spoken in former Soviet states (other than Ukrainian), it will shut down instead of stealing data. It’s not exactly a smoking gun, but it’s pretty suspicious.

After evaluating Telegram channels, Group I-B believes that Godfather is an example of Malware-as-a-Service (MaaS). The creators essentially license the malware to third parties, which can deliver them juicy financial details without the hassle of developing the malware and infrastructure. It targets institutions all over the world, including the US (49 sites), Turkey (31), Spain (30), and Canada (22). If you think you’ve been infected, remove accessibility from all installed apps (usually under Settings > Accessibility) and change your important passwords using a different device.

Now read:

AMD CEO Expected to Announce Zen 4 X3D, RDNA3 GPUs at CES Opening Keynote

The CES trade show takes place in a few weeks, and AMD’s CEO will be delivering the opening keynote. Dr. Lisa Su will take the stage in Vegas, and is expected to deliver the company’s vision for the future. That should include the highly-anticipated Zen 4 CPUs with 3D Cache, as well as more members of the RDNA3 GPU family. The keynote will be live-streamed on Wednesday, Jan. 4, 2023, at 6:30 p.m. from the Venetian.

The most anticipated announcement is Zen 4 CPUs with V-Cache. Although people were generally impressed by Zen 4 CPU performance at launch, the existence of the previous V-Cache CPU kind of crashed the party. AMD conspicuously left that CPU out of its comparisons, seemingly because it was still so competitive against the new architecture. Intel even tried to hide the 5800X3D in its Raptor Lake comparison slides, too. V-Cache CPUs can dramatically boost performance in games, so people are giddy with anticipation about that being applied to Zen 4. The company is expected to announce three V-Cache CPUs, which is a break from tradition. It previously only announced a single X3D CPU and did so at the very end of AM4’s lifecycle. This time around, AMD is facing a serious threat from Intel’s Raptor Lake, which is just as fast and less expensive to upgrade to than AMD’s new CPUs. Therefore, it’s rumored to be moving the announcement up to blunt Intel’s advance.

Dr. Su will likely also be announcing the mainstream lineup for Zen 4. These are the more affordable, non-X versions of its CPUs, such as the 7600, 7700, and so forth. AMD is also anticipated to finally unveil the 8000 series mobile lineup, as well as the previously mentioned Dragon Range mobile CPUs. The company announced their existence on an investor call in May of this year, but we’ve heard nothing since. However, the company promised to offer the highest “thread, core, and cache” amounts for a mobile gaming CPU ever. That’s a lofty promise in light of the fact that Intel is also readying a 24-core mobile chip based on Raptor Lake. AMD will likely also unveil its Phoenix APUs.

AMD has only announced its flagship RDNA3 GPUs so far. Therefore, anticipation is high for the RX 7800 XT, and possibly an RX 7700 XT. Nvidia has also held off on announcing any midrange GPUs. This has led to a market hungry for an affordable GPU using next-gen architecture. Nvidia will probably announce the RTX 4070 Ti at CES, which is the formerly cancelled RTX 4080 Ti. Pricing is TBD, but since it’s Nvidia, it’ll probably be high. This gives AMD an opening to once again undercut Nvidia on price.

“I am excited for the opportunity to deliver a keynote at CES 2023 to highlight the next generation of high-performance and adaptive computing innovations, and products that will push the boundaries on what is possible and play an important role helping solve our most important challenges,” said Dr. Su in the announcement.

Now Read:

Thursday, 22 December 2022

New OpenAI Tool Turns Text Into 3D Models

It’s unusual for an innovation to make equally large waves in the art and artificial intelligence communities, but that’s what OpenAI accomplished with the release of the DALL-E image generator. Simply type in a description, and DALL-E can make it real. The company’s new algorithm has a similar pitch, but instead of making a 2D image, Point-E creates a 3D model of your description.

Point-E doesn’t go right from text to a 3D mesh — it’s actually composed of two different AI models. First, a text-to-image AI processes the prompt to make a standard 2D image. Then, an image-to-3D AI swings into action to turn the flat rendering into a 3D model. So, if you were to ask Point-E to make a traffic cone, it would start with a stripey triangle. It’s up to the second AI to understand that traffic cones are cones to produce the correct 3D shape.

This isn’t a completely novel idea — Google has a tool called DreamFusion that can do something similar. However, DreamFusion was designed to run on a machine with four of Google’s custom TPU v4 AI processors, and it takes that hardware 90 minutes to generate an image. You’re looking at multiple hours of GPU time per image with DreamFusion. Point-E is much faster, and it can run on a computer with a single GPU.

Many of us have had a blast telling DALL-E to make off-the-wall renderings, but Point-E isn’t yet ready for that instant AI gratification. The study explains this is the first step in foundational technology that could eventually become as quick and easy as DALL-E. OpenAI says that Point-E still falls far short of commercial 3D modeling — the results are more akin to a cloud of points.

Point-E models converted to a mesh.

When smoothed and processed, the Point-E models can produce a passable representation of a real object, as seen above. The real leap forward here is the efficiency, which OpenAI claims is “one to two orders of magnitude faster” than existing systems. Perhaps in the future, Point-E will intrude upon the domain of 3D modeling the way DALL-E has in art. Adobe recently decided it would allow AI-generated art in its stock image library, which not all artists like.

If you want to tinker with Point-E, all the code is available on GitHub. However, you’ll need to have Python, as well as some experience with programming and command line tools to make it work. However, the relatively modest hardware requirements mean it’s more accessible than DreamFusion.

Now read: