How Smart Are AI Chips, Really?

November 4, 2019 by Maya Posch 52 Comments

The best part about the term “Artificial Intelligence” is that nobody can really tell you what it exactly means. The main reason for this stems from the term “intelligence”, with definitions ranging from the ability to practice logical reasoning to the ability to perform cognitive tasks or dream up symphonies. When it comes to human intelligence, properties such as self-awareness, complex cognitive feats, and the ability to plan and motivate oneself are generally considered to be defining features. But frankly, what is and isn’t “intelligence” is open to debate.

What isn’t open to debate is that AI is a marketing goldmine. The vagueness has allowed for marketing departments around the world to go all AI-happy, declaring that their product is AI-enabled and insisting that their speech assistant responds ‘intelligently’ to one’s queries. One might begin to believe that we’re on the cusp of a fantastic future inhabited by androids and strong AIs attending to our every whim.

In this article we’ll be looking at the reality behind these claims and ponder humanity’s progress towards becoming a Type I civilization. But this is Hackaday, so we’re also going to dig into the guts of some AI chips, including the Kendryte K210 and see how the hardware of today fits into our Glorious Future. Continue reading “How Smart Are AI Chips, Really?” →

DSP Spreadsheet: Frequency Mixing

November 1, 2019 by Al Williams 18 Comments

Circuit simulation and software workbooks like Matlab and Jupyter are great for being able to build things without a lot of overhead. But these all have some learning curve and often use clever tricks, abstractions, or library calls to obscure what’s really happening. Sometimes it is clearer to build math models in a spreadsheet.

You might think that spreadsheets aren’t built for doing frequency calculation and visualization but you’re wrong. That’s exactly what they’re made for — performing simple but repetative math and helping make sense of the results.

In this installment of the DSP Spreadsheet series, I’m going to talk about two simple yet fundamental things you’ll need to create mathematical models of signals: generating signals and mixing them. Since it is ubiquitous, I’ll use Google Sheets. Most of these examples will work on any spreadsheet, but at least everyone can share a Google Sheets document. Along the way, we’ll see a neat spreadsheet trick I should probably use more often.

Continue reading “DSP Spreadsheet: Frequency Mixing” →

This Week In Security: Project Zero’s IPhone, BBC The Onion, Rooting Androids, And More

November 1, 2019 by Jonathan Bennett 11 Comments

The always interesting Project Zero has a pair of stories revolving around security research itself. The first, from this week, is all about one man’s quest to build a debug iPhone for research. [Brandon Azad] wanted iOS debugging features like single-stepping, turning off certain mitigations, and using the LLDB debugger. While Apple makes debug iPhones, those are rare devices and apparently difficult to get access to.

[Brandon] started looking at the iBoot bootloader, but quickly turned his attention to the debugging facilities baked into the Arm chipset. Between the available XNU source and public Arm documentation, he managed to find and access the CoreSight debug registers, giving him single-step control over a core at a time. By triggering a core halt and then interrupting that core during reset, he was able to disable the code execution protections, giving him essentially everything he was looking for. Accessing this debug interface still requires a kernel level vulnerability, so don’t worry about this research being used maliciously.

The second Google Zero story that caught my eye was published earlier in the month, and is all about finding useful information in unexpected places. Namely, finding debugging symbols in old versions of Adobe Reader. Trying to understand what’s happening under the hood of a running application is challenging when all you have is a decompiler output. Adobe doesn’t ship debug builds of Reader, and has never shipped debug information on Windows. Reader has been around for a long time, and has supported quite a few architectures over the years, and surprisingly quite a few debug builds have been shipped as a result.

How useful could ancient debugging data be? Keep in mind that Adobe changes as little as possible between releases. Some code paradigms, like enums, tend to be rather static as well. Additional elements might be added to the end of the enum, but the existing values are unlikely to change. [Mateusz Jurczyk], the article’s author, then walks us through an example of how to take that data and apply it to figuring out what’s going on with a crash. Continue reading “This Week In Security: Project Zero’s IPhone, BBC The Onion, Rooting Androids, And More” →

Saintcon Badge Is An Enigma No More

October 31, 2019 by Kerry Scharfglass 14 Comments

Through the weekend Twitter has been a-titter with news coming out of Saintcon, the annual security conference in Provo, Utah. Now that the weekend is over we can finally get our hands on full hardware and software sources for the curvy, LED-covered badge we’ve been salivating over and a write up by its creators [compukidmike] and [bashNinja]. Let’s dive in and see what’s waiting!

Design

This year’s badge is designed to represent a single tooth on a single rotor of an Enigma machine. The full function of an Enigma machine is quite complex, but an individual device has three rotors with 26 teeth each (one for each letter) as well as a keypad for input and a character display to show each enciphered letter. For reference, the back of the badge has a handy diagram of a badge’s place in the Enigma system.

Reminiscent of the WWII device which the badge design recalls, each unit includes a full QWERTZ keyboard (with labeled keys!) and RGB “lampboard” for individual character output, but unlike the original there’s also a curved 16 x 64 RGB LED display made from those beguiling little ~1mm x 1mm LEDs. All in, the device includes 1051 LEDs! Combined with the unusually non-rectilinear shape of the badge and the Enigma-style Saintcon logo it makes for an attractive, cohesive look.

Continue reading “Saintcon Badge Is An Enigma No More” →

Linux’s Marketing Problem

October 31, 2019 by Bryan Cockfield 265 Comments

The cult classic movie Office Space is a scathing critique of life for software engineers in a cubicle farm, and it did get a lot of things right even if it didn’t always mean to. One of those is the character of Tom Smykowski whose job is to “deal with the customers so the engineers don’t have to”. The movie treats Tom and his job as a punchline in a way, but his role is actually very important for most real businesses that rely on engineers or programmers for their core products.

Engineers can have difficulty relating to customers, and often don’t have the time (or even willingness) to handle the logistics of interacting with them in the first place. Customers may get frustrated understanding engineers or communicating their ideas clearly to them. A person like Tom Smykowski is often necessary to bridge the gap and smooth out the rough edges on both sides, but in the Linux world there are very few Toms to rely on. The customers, or users, have to deal directly with the engineers in many situations, and it’s not working out very well for either group. Linux has a marketing problem, and it needs a marketing solution if it ever wants to increase its market share in the PC realm. Continue reading “Linux’s Marketing Problem” →

RPi4: Now Overclocked, Net-Booted, And Power-Sipping

October 30, 2019 by Jenny List 58 Comments

It has now been a few months since the launch of the Raspberry Pi 4, and it would only be fair to describe the launch as “rocky”. While significantly faster than the Pi 3 on paper, its propensity for overheating would end up throttling down the CPU clock even with the plethora of aftermarket heatsinks and fans. The Raspberry Pi folks have been working on solutions to these teething troubles, and they have now released a bunch of updates in the form of a new bootloader, that lets the Pi 4 live up to its promise. (UPDATE: Here’s the download page and release notes)

The real meat of the update comes in an implementation of a low power mode for the USB hub. It turns out that the main source of heat on the SoC wasn’t the CPU, but the USB. Fixing the USB power consumption means that you can run the processor cool at stock speeds, and it can even be overclocked now.

There is also a new tool for updating the Pi bootloader, rpi-eeprom, that allows automatic updates for Pi 4 owners. The big change is that booting the Pi 4 over the network or an attached USB device is now a possibility, which is a must if you’re installing the Pi permanently. There are some fixes that caused problems with certain HATs, in which the Pi 4’s 3.3 V line was cycled during a reboot.

With a device as complex as a Raspberry Pi it comes as no surprise that it might ship with a few teething troubles. We’ve already covered some surrounding the USB-C power, for example. And the overheating. Where the Pi people consistently deliver though is in terms of support, both official and from the community, and we’re very pleased to see them come through in this case too.

5G Is For Robots

October 30, 2019 by Gerrit Coetzee 45 Comments

Ecclesiastes 1:9 reads “What has been will be again, what has done will be done again; there is nothing new under the sun.” Or in other words, 5G is mostly marketing nonsense; like 4G, 3G, and 2G was before it. Let’s not forget LTE, 4G LTE, Advance 4G, and Edge.

Just a normal everyday antenna array in a Seattle parking garage.

Technically, 5G means that providers could, if they wanted to, install some EHF antennas; the same kind we’ve been using forever to do point to point microwave internet in cities. These frequencies are too lazy to pass through a wall, so we’d have to install these antennas in a grid at ground level. The promised result is that we’ll all get slightly lower latency tiered internet connections that won’t live up to the hype at all. From a customer perspective, about the only thing it will do is let us hit the 8Gb ceiling twice as faster on our “unlimited” plans before they throttle us. It might be nice on a laptop, but it would be a historically ridiculous assumption that Verizon is going to let us tether devices to their shiny new network without charging us a million Yen for the privilege.

So, what’s the deal? From a practical standpoint we’ve already maxed out what a phone needs. For example, here’s a dirty secret of the phone world: you can’t tell the difference between 1080p and 720p video on a tiny screen. I know of more than one company where the 1080p on their app really means 640 or 720 displayed on the device and 1080p is recorded on the cloud somewhere for download. Not a single user has noticed or complained. Oh, maybe if you’re looking hard you can feel that one picture is sharper than the other, but past that what are you doing? Likewise, what’s the point of 60fps 8k video on a phone? Or even a laptop for that matter?

Are we really going to max out a mobile webpage? Since our device’s ability to present information exceeds our ability to process it, is there a theoretical maximum to the size of an app? Even if we had Gbit internet to every phone in the world, from a user standpoint it would be a marginal improvement at best. Unless you’re a professional mobile game player (is that a thing yet?) latency is meaningless to you. The buffer buffs the experience until it shines.

So why should we care about billion dollar corporations racing to have the best network for sending low resolution advertising gifs to our disctracto cubes? Because 5G is for robots.

Continue reading “5G Is For Robots” →