Tech Brew Ride Home - Tue 10/12 – Twitter Lets Your Do Your Own Shadow Banning

Episode Date: October 12, 2021

Twitter lets you remove followers without having to block them. Magic Leap is back with a fresh $500 million dollars. Canva gets into video while 1Password wants to make it easier to share your password. Microsoft and Nvidia claim to have trained the largest AI language model yet. And when it comes to AI, is China about to dominate everyone else, or are they only focused on one thing? Sponsors: Masterworks.io/ride Gainful.com/techmeme Links: Twitter's tool for removing unwanted followers arrives for web users (Engadget) Facebook whistleblower to brief Facebook Oversight Board, U.K. parliament (Axios) A second Facebook whistleblower says she's willing to testify before Congress and that she's shared documents with a US law agency (Insider) Seven years after raising $542M at a $2B valuation, Magic Leap raises $500M at a $2B valuation (TechCrunch) Canva is getting into video (TechCrunch) 1Password's new feature lets you safely share passwords using just a link (Engadget) Microsoft said it mitigated a 2.4 Tbps DDoS attack, the largest ever (The Record) Microsoft and Nvidia team up to train one of the world’s largest language models (VentureBeat) US has already lost AI fight to China, says ex-Pentagon software chief (Financial Times) Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the TechMemebrite home for Tuesday, October 12th, 2021. I'm Brian McCullough today. Twitter lets you remove followers without having to block them. Magic Leap is back with a fresh $500 million. Canva gets into video while one password wants to make it easier for you to actually share your password.

Microsoft and InVIDIA claim to have trained the largest AI language model yet. And when it comes to AI, is China about to dominate everyone else or are they only focused on one thing? Here's what you miss today in the world of tech. Twitter now lets web users remove followers without having to block them after testing the option for over a month. Notice, though, that I said web users, quoting Engadgett. All Twitter users can now remove a follower without having to block them. To quietly stop someone from seeing your tweets in their feed, go to the followers tab on your profile, click the three dot menu next to the user in question, and select the remove this follower option. This is part of Twitter's efforts to reduce harassment on the platform,

Blocking someone you don't want to follow you could lead to retaliation from that person via their allies or their secondary accounts after they find out. But cutting them in this fashion and muting them will mean they're none the wiser that they're out of the loop. This method won't prevent someone you boot from your followers list from seeing your public tweets. Only blocking them or making your account private will do that, end quote. I sort of been letting the continuing Facebook controversy royal on without talking about it in much detail, because there hasn't been much new in the way of details, just a bunch of back-and-forth PR battling. Though this is interesting, whistleblower Francis Howgan has accepted an invitation to brief Facebook's oversight board, which she claims Facebook has

lied to repeatedly, quoting Axios. The oversight board said in a statement it wants to, quote, gather information that can help push for greater transparency and accountability, end quote. Hougain's busy schedule this month will include appearing before the UK Parliament on October 25th to give evidence to the Joint Committee considering the online safety bill, legislation geared toward regulating social media companies, end quote. And apparently a second Facebook whistleblower says she's willing to testify before Congress and that she's shared documents with a U.S. law agency. This is from insider, quote, Sophie Zhang, a former Facebook data scientist who went public with her criticisms of the company in September 2020, has told CNN she is willing to testify before Congress.

Zhang also said on Twitter on Sunday that she had provided a U.S. law enforcement agency with, quote, detailed documentation regarding potential criminal violations, end quote. When asked by CNN, Zhang did not say which agency she gave documents to, and FBI spokesperson declined to comment when contacted by CNN. If Congress wishes for me to testify, I will fulfill my civic duty, as I've publicly stated for the past half year, Zhang said in a tweet Monday that linked to her CNN interview. Speaking to CNN, Zhang said she was encouraged by the apparent bipartisan support for action against Facebook following Francis Hogan, another Facebook whistleblower, testifying about

children's safety on Facebook and Instagram in a congressional hearing on October 5th. Zhang was fired from Facebook in August 2020, but before she left, she posted a 7,800-word memo detailing how she believed the company allowed authoritarian regimes around the world to manipulate its platform. I have blood on my hands, Zhang wrote in the memo, which was obtained by BuzzFeed. Zhang wrote that she was officially being fired for, quote, poor performance, end quote. 2019 called and they want some of their headlines back. If we work is back, why not Magic Leap? If you'll recall, when we last left Magic Leap, it seemed it was being sold for parts,

or at least pivoting towards some sort of enterprise play. This is after raising a metric ton of venture capital money on the back of demos that people swore was the most jaw-dropping they had ever seen. And yet, the actual product, when it arrived, seemed to impress nobody. Well, Magic Leap has raised a fresh $500 million and says its second-gen-gen-a-R headset will arrive in 2022 with a slimmer form factor and twice the field of view. This came in a CNBC interview with CEO Peggy Johnson, Quoting TechCrunch. The augmented reality startup announced today that they have raised $500 million at a $2 billion valuation from existing investors.

The round echoes the terms of an October 2014 raise where Magic Leap raised $542 million at a reported $2 billion valuation. Quite a bit has happened in the meantime. Curiously, Magic Leap decided not to actually disclose any of the specific investors participating in this latest fundraise. At this point, the company has raised $3.5 billion. and total funding, according to Crunch Base, meaning that most of the investors they've brought in haven't fared too well thus far. The latest valuation is a far cry from the company's $6.7 billion valuation they reached in 2019, but for a startup that nearly shut down last year, it could be a lot worse. Magic Leap fired a healthy chunk of its staff and was forced to raise hundreds of millions

at slashed valuation terms last year. The startup also replaced its founder Ronnie Abavitz with former Microsoft VP, Peggy Johnson as CEO. The company has been trying to turn things around by revamping its messaging with a focus wholly on enterprise audiences after years of trying to compete with Microsoft for military contracts and failing, and trying to compete with Microsoft for enterprise clients with some success, all while attempting to also stay on consumer radars by bankrolling expensive games for their very expensive headset, end quote. Quick news you can maybe use segment, tools that you can maybe make use of right now. Canva, for example, has announced Canva Video, a suite of video editing tools available as part of its free tier. Yes, Canva wants to do for video

what it's done for all other mediums of design, quoting TechCrunch. Launched in 2013, Canva emerged during the design renaissance around the same time that Sketch, Figma, and Invision were crashing the Adobe Party. Unlike the others, Canva focused on democratization of design, creating a UI that allowed anyone to make just about anything, from slide decks to t-shirts to social media assets. Today, Canva continues on that trajectory in the world of video. The Canva video software offers a scene-focused interface, meaning that users can break down their video into easy-to-manage pieces. From there, they can trim scenes, coming soon, manage multiple audio tracks, and use in-editor video previews to check their work. Canva is also offering video

templates to simplify the process for users based on their end goals. For example, folks making a TikTok video are designing and editing in a different format than folks making Twitch banners. Moreover, Canva Video is multiplayer, allowing users to edit in the same scene-based editor without having to download, re-upload, etc., end quote. Which is all pretty impressive that you can get all that in the free tier. And also, one password has launched, PST, which lets users share login credentials with just a link. PST is actually a clever acronym, standing for password-secure sharing tool or PSST,

quoting and gadget. Based on research, the company conducted most workers reuse corporate credentials. Of those people, 36% admitted to sharing logins with other workers and clients via insecure channels including email, chat apps, spreadsheets, documents, and texts. In another survey, one password found that most families share passwords between members, and also use insecure methods such as writing them down and messaging them to each other. PST allows you to share credentials with anybody by generating a link with the information that other people can access, even if they don't have a one-password account.

You can choose how long the link remains valid from an hour to 30 days, after which it will automatically expire. You can also choose to share the credentials with anyone who has the link or with specific people only. If you choose the latter, the service will require the recipient to confirm their identity by verifying their email address before giving them access, end quote. Microsoft says it has mitigated a 2.4 terabytes per second DDoS attack on an Azure customer at the end of August, breaking AWS's previous record for denial of service attacks, mitigating a 2.3 terabytes per second attack in February 2020, quoting the record. Amir DeHan, senior program manager for Azure networking, said the attack was carried out using

a botnet primarily located across the Asia-Pacific region, such as Malaysia, Vietnam, Taiwan, Japan, and China, as well as the United States. DeHan identified the target of the attack only as an Azure customer in Europe. The Microsoft execs said the record-breaking attack came in three short waves in the span of 10 minutes with the first at 2.4 terabytes per second and the second at 0.55 terabytes per second, while the third came in at 1.7. DeHan said Microsoft successfully mitigated the attack without Azure going down, end quote. Meanwhile, Microsoft and Nvidia claimed to have trained the largest and most capable AI language model yet containing 530 billion parameters, quoting Venture Beat. The successor to the company's Turing NLG-17B and Megatron-LM models, MTNLP contains 530 billion

parameters and achieves unmatched accuracy and a broad set of natural language tasks. Microsoft and Nvidia say, including reading comprehension, common sense reasoning, and natural language inferences. In machine learning parameters are the part of the model that's learned from historical training data. Generally speaking, in the language domain, the correlation between the number of parameters and sophistication has held up remarkably well. Language models with large numbers of parameters, more data and more training time, have been shown to acquire a richer, more nuanced understanding of language, for example, gaining the ability to summarize books and even complete programming code. To train MTNLG, Microsoft and NVIDIA say that they created a training

data set with 270 billion tokens from English language websites. Tocens, a way of separating pieces of text into smaller units in natural language, can either be words, characters, or parts of words. Like all AI models, MTNLP had to train by ingesting a set of examples to learn patterns among data points like grammatical and syntactical rules. The dataset largely came from the pile and 835 gigabyte collection of 22 smaller datasets created by the open source AI research effort, E. Luther AI. The pile spans academic sources, communities like stack exchange and Wikipedia, code repositories like GitHub and more, which Microsoft and NVIDIA say they curated and combined with filtered snapshots of the Common Crawl, a large collection of web pages including news stories and

social media posts, end quote. Finally today, ex-Penagon software chief, Nicholas Chilin, who recently resigned from his position, says China is headed for global dominance due to its advances in AI, machine learning, and cyber capabilities, quoting the Financial Times. In his first interview since leaving the post at the Department of Defense a week ago, Nicholas Chilin told the Financial Times that the failure of the U.S. to respond to Chinese cyber and other threats was putting his children's future at risk, quote, we have no competing fighting chance against China in 15 to 20 years. Right now, it's already a done deal. It is already over, in my opinion, he said, adding there was, quote, good reason to be angry, end quote. Chilin 37, who spent three years on a Pentagon-wide effort

to boost cybersecurity and as first chief software officer of the U.S. Air Force, said Beijing is heading for global dominance because of its advances in artificial intelligence, machine learning, and cyber capabilities. He argued these emerging technologies were far more critical to America's future than hardware such as big-budget fifth-generation fighter jets, such as the F-35. Whether it takes a war or not is kind of anecdotal, he said, arguing China was set to dominate the future of the world controlling everything from media narratives to geopolitics. He added U.S. cyber defenses and some government departments were at, quote, kindergarten level. He also blamed the reluctance of Google to work with the U.S. Defense Department on AI and extensive debates over AI ethics for slowing the U.S. down.

By contrast, he said Chinese companies are obliged to work with Beijing and were making massive investments into AI without regard to ethics. Chilin says he plans to testify to Congress about the Chinese cyber threat to U.S. supremacy, including in classified briefings over the coming weeks. He acknowledged the U.S. still outspends China by three times on defense, but said the extra cash was immaterial because U.S. procurement costs were so high and spent in the wrong areas, while bureaucracy and overregulation stood in the way of much-needed change at the Pentagon, end quote. Big if true, as they say, but at the same time, there's this. In Bloomberg opinion this morning tech columnist Tim Culpin says, despite fears in the West, China's alleged AI prowess is mostly

limited to computer vision, which is useful for domestic control of citizens because priorities, the U.S.'s AI tech is far broader. Quote, on paper, the U.S. and China appear neck and neck in artificial intelligence. China leads in the share of journal citations, helped by the fact that it also publishes more, while the U.S. is far ahead in the more qualitative metric of cited conference papers, according to a recent report compiled by Stanford University. So while the world's most populous country is an AI superpower, investors and China watchers shouldn't put too much stock in the notion that its position is unassailable or that the U.S. is weaker. By miscalculating the others' abilities, both superpowers risk overestimating their adversary's strengths and overcompensating

in a way that could lead to a Cold War-style AI arms race. In fact, China's expertise is somewhat limited in scope. Artificial intelligence encompasses many subfields, including machine learning, robotics, natural language processing, and computer vision. The U.S. has a broad toolkit that's deployed across each of these disciplines and used around the world. China, by contrast, excels mostly in computer vision, an area that helps Beijing build out its surveillance state. More than half the papers last year published by Chinese teams were in the disciplines most used in surveillance, computer vision, and pattern recognition, according to data on research papers compiled by Microsoft Academic Graph and analyzed by Chinese media outlet sixth tone.

Compared with the U.S., UK, Germany, and India, a smaller proportion of its papers are on machine learning, natural language processing, and robotics. Revenue for AI computer vision companies in China, meanwhile, is largely linked to governments and their desire to monitor people. Sense Time, which is planning a Hong Kong public offering, got almost half its revenue in the first six months of the year from its smart city business. Rival Megvi technology, which swapped its earlier listing plans in the city for an upcoming Shanghai debut, relies even more heavily on local government contracts with 64% of its business coming from a similar product suite it calls city IOT solutions. These projects are largely based on

setting up cameras and sensors around a city to track people and vehicles, then using AI software to keep tabs on their movement. AI technology in the U.S. has many more applications. The country leads in machine learning, a branch of AI that leverages data and algorithms to learn and improve accuracy. Most of the world's leading and widely adopted frameworks are developed in America by companies such as Google, Microsoft, and Facebook, as well as the University of California, Berkeley. The best natural language processing engines used to understand process and analyze written text also hail from U.S. Tech Giants, Microsoft, Google, Amazon, and IBM. These tools are used globally for voice recognition and smart speakers to run chatbots in customer service, improve translation

and search results, or detect fraud in the financial industry. U.S. companies are also leaders in enterprise deployments, where the recent surge in work from home usage has driven demand for cloud computing and networking technologies, and it's a world-beater in process automation, which is used to replace staff in call centers and customer service, end quote. It's almost as if one country's technology is being developed through the lens of and in service of, one party authoritarianism, and the other country is developing its tech through the lens and in service of capitalism. As I was editing this, Apple announced a virtual event titled Unleashed that will be held on Monday, October 18th at 10 a.m. Pacific time, 1 p.m. Eastern, where new MacBook Pros with M1X chips are

rumored. So think of this as Christmas coming early for Brian, or very, very late, considering how very, very long I've had to wait for a MacBook Pro that I actually want to use again. Fingers crossed. Talk to you tomorrow.