Google Archive

Google’s plan is that, during a webpage transaction, the web server could require you to pass an “environment attestation” test before you get any data. At this point your browser would contact a “third-party” attestation server, and you would need to pass some kind of test. If you passed, you would get a signed “IntegrityToken” that verifies your environment is unmodified and points to the content you wanted unlocked. You bring this back to the web server, and if the server trusts the attestation company, you get the content unlocked and finally get a response with the data you wanted. The web mercilessly mocked this idiotic proposal over the weekend, and rightfully so. This is an unadulterated, transparent attempt at locking down the web with DRM-like nonsense just to serve more targeted ads that you can’t block. This must not make its way into any browser or onto any server in any way, shape, or form. The less attention we give to this drivel, the better.

The contractors are the invisible backend of the generative AI boom that’s hyped to change everything. Chatbots like Bard use computer intelligence to respond almost instantly to a range of queries spanning all of human knowledge and creativity. But to improve those responses so they can be reliably delivered again and again, tech companies rely on actual people who review the answers, provide feedback on mistakes and weed out any inklings of bias. It’s an increasingly thankless job. Six current Google contract workers said that as the company entered a AI arms race with rival OpenAI over the past year, the size of their workload and complexity of their tasks increased. Without specific expertise, they were trusted to assess answers in subjects ranging from medication doses to state laws. Documents shared with Bloomberg show convoluted instructions that workers must apply to tasks with deadlines for auditing answers that can be as short as three minutes. That’s the reality of “artificial intelligence” – the same reality it always seems to be in Silicon Valley: thousands and thousands of exploited workers behind the scenes running around like ants keeping the illusion of futurism alive for meager pay.

Eight years after Google Domains launched, and a little more than a year after it graduated out of beta, Google is “winding down following a transition period,” as part of “efforts to sharpen our focus.” That’s corporate-ese for “We need to keep cost-cutting, so we’re selling this business we just finished shaping up to Squarespace.” I have two domains over at Google Domains. I doubt Squarespace’s UI is going to be as nice and easy to understand as Google’s is.

We’re bringing a new mid-tier compiler to Chrome. Maglev is a just-in-time compiler that can quickly generate performant machine code for all relevant functions within the first one-hundredth of a second. It reduces overall CPU time to compile code while also saving battery life. Our measurements show Maglev has provided a 7.5 percent improvement on Jetstream and a 5 percent improvement in Speedometer. Maglev will start rolling out in Chrome version 114, which begins release on June 5. Let’s hope making benchmarks run faster also makes actual websites load faster.

On a support page, Google details the full list of 180 countries in which Bard is now available. This includes countries all over the globe, but very noticeably not any countries that are a part of the European Union. It’s a big absence from what is otherwise a global expansion for Google’s AI. The reason why isn’t officially stated by Google, but it seems reasonable to believe that it’s related to GDPR. Just last month, Italy briefly banned ChatGPT over similar concerns that the AI couldn’t comply with the regulations. Google also slyly hints this might be the case saying that further Bard expansions will be made “consistent with local regulations.” In other words, Bard probably does things that run afoul of the stricter privacy regulations in the EU. Make of that what you will.

Have you ever found yourself in this position? You see an image on a website, in your feed, or in a message from a friend — and you think, “this doesn’t feel quite right.” Is the image being shown in the right context? Has it been manipulated or faked? Where did it come from? When you’re trying to figure out if a piece of information or an image is reliable, having the full story is key. That’s why we’re expanding our ongoing work in information literacy to include more visual literacy and help people quickly and easily assess the context and credibility of images. In the coming months, we’re launching a new tool called About this image. This is a great idea, and I hope it works as intended. While I doubt it’ll be perfect, it’ll make it much easier to quickly verify where an image came from, just how genuine or fake it is, if it’s been edited, and more. It’s not giving a simple “yay” or “nay”, but instead gives the user the data it can then use to make their own informed decision. This is the kind of stuff Google should be doing.

Google I/O, Google’s developer conference, started today, and there has been a deluge of news coming out of the advertising giant. I do not intend to cover every single bit of I/O news, instead choosing to focus one some of the more interesting bits and pieces. In the coming weeks, when you search for something that might benefit from the experiences of others, you may see a Perspectives filter appear at the top of search results. Tap the filter, and you’ll exclusively see long- and short-form videos, images and written posts that people have shared on discussion boards, Q&A sites and social media platforms. We’ll also show more details about the creators of this content, such as their name, profile photo or information about the popularity of their content. Basically, this is a “remove SEO garbage” button. Whenever I need to find some answer to a tech issue or see if other people are experiencing a bug, regular Google search is entirely useless, as the results are overflowing with useless SEO/AI garbage, so I do what a lot of us do: append “reddit” to our queries to get content from real people. With this new Perspectives filter, Google seems to finally acknowledge that their regular search results are useless, and that what users really want is genuine results written by normal humans. I really hope this works as advertised.

And yet, 25 years on, Google Search faces a series of interlocking AI-related challenges that together represent an existential threat to Google itself.  The first is a problem of Google’s own making: the SEO monster has eaten the user experience of search from the inside out. Searching the web for information is an increasingly user-hostile experience, an arbitrage racket run by search-optimized content sharks running an ever-changing series of monetization hustles with no regard for anything but collecting the most pennies at the biggest scale. AI-powered content farms focused on high-value search terms like heat-seeking missiles are already here; Google is only now catching up, and its response to them will change how it sends traffic around the web in momentous ways. That leads to the second problem, which is that chat-based search tools like Microsoft’s Bing and Google’s own Bard represent something that feels like the future of search, without any of the corresponding business models or revenue that Google has built up over the past 25 years. If Google Search continues to degrade in quality, people will switch to better options — a switch that venture-backed startups and well-funded competitors like Microsoft are more than happy to subsidize in search of growth, but which directly impacts Google’s bottom line. At the same time, Google’s paying tens of billions annually to device makers like Apple and Samsung to be the default search engine on phones. Those deals are up for renewal, and there will be no pity for Google’s margins in these negotiations. Search on the web is in a terrible state right now. Searching for anything on Google is a horrible experience, with results riddled with ads and an endless stream of SEO’d garbage content of low to no quality. Alternatives, such as DuckDuckGo, aren’t much better, and tend to promote garbage anti-science and fascist nonsense if you’re not careful enough. At this point I just don’t know what to use to find stuff on the web, and tend to just go straight to sites that I think have the best odds of containing a relevant result (e.g. going straight to Reddit when dealing with some obscure bug or software issue). I know there are even smaller competitors, but I don’t hold high hopes they can offer the same breadth as Google once did, or even DuckDuckGo sometimes does now. It’s not looking pretty out there.

Google has 175,000+ capable and well-compensated employees who get very little done quarter over quarter, year over year. Like mice, they are trapped in a maze of approvals, launch processes, legal reviews, performance reviews, exec reviews, documents, meetings, bug reports, triage, OKRs, H1 plans followed by H2 plans, all-hands summits, and inevitable reorgs. The mice are regularly fed their “cheese” (promotions, bonuses, fancy food, fancier perks) and despite many wanting to experience personal satisfaction and impact from their work, the system trains them to quell these inappropriate desires and learn what it actually means to be “Googley” — just don’t rock the boat. As Deepak Malhotra put it in his excellent business fable, at some point the problem is no longer that the mouse is in a maze. The problem is that “the maze is in the mouse”. I have never worked at any company – other than the hardware store for 7-8 years when I was a teenager and during university – so I have no idea if this is uncommon, but this sounds like my personal version of hell. No wonder Google has such a massive graveyard.

The Justice Department and a group of eight states sued Google on Tuesday, accusing it of illegally abusing a monopoly over the technology that powers online advertising, in the agency’s first antitrust lawsuit against a tech giant under President Biden and an escalation in legal pressure on one of the world’s biggest internet companies. The lawsuit said Google had “corrupted legitimate competition in the ad tech industry by engaging in a systematic campaign to seize control of the wide swath of high-tech tools used by publishers, advertisers and brokers to facilitate digital advertising.” More of this, please.

To begin collaborating with others, we’ve open sourced several components for our secure operating system, called KataOS, on GitHub, as well as partnered with Antmicro on their Renode simulator and related frameworks. As the foundation for this new operating system, we chose seL4 as the microkernel because it puts security front and center; it is mathematically proven secure, with guaranteed confidentiality, integrity, and availability. Through the seL4 CAmkES framework, we’re also able to provide statically-defined and analyzable system components. KataOS provides a verifiably-secure platform that protects the user’s privacy because it is logically impossible for applications to breach the kernel’s hardware security protections and the system components are verifiably secure. KataOS is also implemented almost entirely in Rust, which provides a strong starting point for software security, since it eliminates entire classes of bugs, such as off-by-one errors and buffer overflows. Another new open source operating system by Google. This time, it seems almost entirely focused on embedded machine learning applications, so it’s definitely a bit outside of my wheel house.

Swider wasn’t the only Stadia developer blindsided by Google’s late September announcement that the streaming gaming service would be shutting down next January. Game makers who talked to Ars (and some who shared their surprise on social media) all said they had no indication of Google’s shutdown plans before the public announcement. “During correspondence , we are exchanging emails—nothing showed us it could be the end of Stadia,” Swider said. What a shitshow.

A few years ago, we also launched a consumer gaming service, Stadia. And while Stadia’s approach to streaming games for consumers was built on a strong technology foundation, it hasn’t gained the traction with users that we expected so we’ve made the difficult decision to begin winding down our Stadia streaming service. We’re grateful to the dedicated Stadia players that have been with us from the start. We will be refunding all Stadia hardware purchases made through the Google Store, and all game and add-on content purchases made through the Stadia store. Players will continue to have access to their games library and play through January 18, 2023 so they can complete final play sessions. We expect to have the majority of refunds completed by mid-January, 2023. Another Google product announced with much fanfare is shutting down, as many, many people expected it would be. It seems Google is at least handling the refunds properly, and I hope the Stadia controllers can still be used with other platforms so they don’t turn into e-waste. Another one for the graveyard.

Roughly seven years ago, Partha Ranganathan realized Moore’s law was dead. That was a pretty big problem for the Google engineering vice president: He had come to expect chip performance to double every 18 months without cost increases and had helped organize purchasing plans for the tens of billions of dollars Google spends on computing infrastructure each year around that idea. But now Ranganathan was getting a chip twice as good every four years, and it looked like that gap was going to stretch out even further in the not-too-distant future. So he and Google decided to do something about it. The company had already committed hundreds of millions of dollars to design its own custom chips for AI, called tensor processing units, or TPUs. Google has now launched more than four generations of the TPU, and the technology has given the company’s AI efforts a leg up over its rivals. Google uses all kinds of custom hardware throughout its operations, but you rarely hear about it. This article provides some insight into the custom hardware Google uses for YouTube transcoding.