Issue 154
š¦¹šø Multinational firm loses $25 million in deepfake scam. Swarms of AI killer robots are the future of war. 17 mysteries that researchers still havenāt solved.
Dear Bizarro Readers,
Welcome to the first of two April editions of Bizarro Devs. As usual, weāve got a nice lineup of stories that touch upon a variety of brain-tickling topics. Thereās a recent cybercrime tale out of Hong Kong, an article that looks ahead to what humanity might be facing in the near future, and finally, a fun post that will make you reflect on some unique, unexplained mysteries. Of course we also have the usual mix of tools, stats, and a new code-related challenge for the Python lovers out there.
Below are the top stories:
š¦¹šø Multinational Firm Loses $25 Million in Deepfake Scam
š¤šŖ Swarms of AI Killer Robots Are the Future of War
š¤š 17 Mysteries That Researchers Still Havenāt Solved
Letās get started!
š° From the Newsroom
š¦¹šø Multinational Firm Loses $25 Million in Deepfake Scam
Imagine being the employee responsible for losing $25 million dollars? Thatās gotta take a psychological toll on anybody. The thing is, with the rise of deepfake technology and its increasingly higher level of sophistication, it could happen to anybody. In this particular situation, it happened to a finance worker in Hong Kong. Hereās how it went down:
The elaborate heist began when the employee received a message that was supposedly from the companyās UK-based chief financial officer. The message mentioned the need for a āsecret transactionā to be carried out. This initially made the worker think it was a phishing email.
Phase two of the scam is when the deepfake technology came into play. The fraudsters correctly predicted he would be suspicious of the request, so they leveraged his co-workers to assuage his doubts.
He was invited to attend a video call with his colleagues to discuss the transfer - except that unbeknownst to him, all of his colleagues were, in fact, AI-generated simulations. Believing everyone else on the call was real, the employee ended up transferring a total of $200 million Hong Kong dollars ā about $25.6 million USD.
Reading this story made me feel a lot of sympathy for this person. It also made me reflect on how much crazier things are going to get with these deepfakes, and how it mirrors whatās been happening with AI-generated text content. ChatGPT (and similar competitors) arenāt even two years old yet, and weāre already drowning in a sea of their output. It feels like at some point all of the tech we use is just going to dissolve into a pile of AI mush. I think we should start a new slogan: make tech human again.
š¤šŖ Swarms of AI Killer Robots Are the Future of War
Drones of all kinds are now a very active part of modern warfare. Despite their variety, almost all of them have one thing in common: theyāre controlled by humans. A notable exception is Israelās drones. Although a human does still make the final attack decision, the Israelis have begun relying on AI to tell them which targets to strike. Author and professor, Michael T. Klare, argues that Israelās approach is going to be adopted by other militaries, with increasingly more agency being given to the AI.
Klare states that given the work on artificial intelligence and robot weaponry that the major powers have already begun, that in the future weāll see an array of self-guided autonomous weapons systems. Essentially, these will be combat drones capable of applying lethal force independently of any human oversight.
In a nod to the global concern over this, the U.N. General Assembly held its first discussion about autonomous weapons last October, and itās planning a full-scale review of the topic this coming fall. The primary concern is whether these machines will be empowered to take a human life without human oversight.
Military strategists, particularly those in the U.S., disregard these concerns and are pushing ahead for a future where they envision āswarmsā of AI-powered killing machines making real-time battlefield decisions by coordinating with each other in an algorithmic network thatās outside the bounds of human processing capability.
Last year, Antony Loewenstein, an independent journalist, best-selling author, and filmmaker released a book titled āThe Palestine Laboratory: How Israel Exports the Technology of Occupation Around the Worldā where he describes how Israelās weapons and technology manufacturers use the Palestinian Territories as testing grounds for their weaponry and spying gadgets so they can then market them around the world as ābattle tested.ā If Gaza is a compartmentalized glimpse of whatās in store for the rest of humanity, then we should all be concerned.
š¤š 17 Mysteries That Researchers Still Havenāt Solved
Scientific research has enabled us to understand a lot about the world around us, ourselves, other animals, plants of all kinds, and more. However, as much as weāve learned, there are still those lingering āmysteries of the universeā that keep us searching for answers. This roundup includes 17 such mysteries. Iāll share three of my favorites from the list below:
How does sound become hearing? We have technology that can show us sound waves, and we know that we have parts inside of our ears that process those sound waves, but then how does our brain process them?
Is there anything alive in the human poop left on the moon? During the Apollo moon missions, astronauts went to the moon and, to save weight for returning to Earth, they left 96 bags of their own waste behind. Were the microbes inside their poop resilient enough to survive?
Was there an advanced civilization on Earth before humans? Is it possible that someday we might find evidence, buried deep within the Earthās crust, of an advanced non-human civilization that lived hundreds of millions of years ago?
Honestly, all of the questions were interesting to think about. Plus each one has a little write up next to it with some related information to get your brain moving. It was tough to narrow it down to only three, and if you asked me a week from now, I might even pick a different three. In short, go read the article for yourself. Youāll enjoy it. I promise. š
āļø Ten Must See Links of the Month
In 2026, NASA will send Artemis 3 to the moon. It will be the first human-staffed expedition to the lunar surface since Apollo 17 in December of 1972. History aside, one of the most interesting aspects of the astronautsā mission will be to set up a greenhouse on the moon - using plants they will bring from Earth.
Inspired by self-healing tissue in the human body, researchers at Drexel University developed a self-repairing concrete material (i.e., cement) that they call BioFiber. BioFiber can heal itself, stop cracks from growing wider, and remain intact inside the concrete when there arenāt any cracks.
Last year, the New York City government launched an AI-powered chatbot to help New Yorkers with starting and operating a business in the city.Ā Itās now being reported that this chatbot has been giving out false information and sometimes even advice that is outright illegal.
If you were impressed last month by Open AIās teaser clips from its new video generator Sora, then you might enjoy reading about where AI moviemaking is headed.
Governments on both sides of the Atlantic have recently made power moves in an effort to assert some kind of control over the booming AI industry. Lawmakers in the European Union passed the EU AI Act, while in the U.S., President Biden has ordered every US government agency to appoint a Chief AI Officer.
One of the most common elements of action movies is the high-speed car chase scene. Those car chases also happen in real life, but we may be seeing less of them - in the U.S. at least. Some police departments there have begun using vehicle-mounted launchers that fire sticky, GPS-enabled projectiles at vehicles. This reduces the need to directly pursue fleeing suspects.
Scammers have been using AI to create fake obituary notice websites that mimic real ones in detail and sentiment. Visitors to these sites have been falsely led to believe that theyāre paying respects to a deceased individual. Instead, the scammers have used the sites to phish for personal data, solicit donations, and infect visitorsā devices with malware.
š½ļøšļø When it comes to network security and encryption, randomness is crucial. Without it, an attacker can eventually figure out the keys that a computer uses to encrypt data - and decrypt them. The problem is that computers are designed to give predictable, logical outputs based on a given input. Cloudflare has developed a truly creative solution for this. It almost seems fake (itās not).
Curious to know about all of the latest developments in CSS? This tutorial covers the newest CSS features for Spring 2024, offering insights and practical examples to help you stay up to date.
I recently interviewed two of my colleagues, both of whom are software engineers, about how they broke into the industry, what their day-to-day is like, and more. One of them works 100% remotely, while the other works hybrid-style, with some of his days in the office and some at home. If youāve ever wondered what itās like to be a software engineer, then this is an insightful read.
š¤ Itās How They Said It
"I call it the 'dot AIā bubble, and it hasnāt even started yet."
ā AI Emad Mostaque, recently ousted CEO of Stability (an AI company), speaking about the looming fallout from the massive amounts of money pouring into AI.
š§® The Numbers Game
$176,500 was the average bonus for Wall Street employees last year, down 2% from $180,000 in 2022. This decline occurred as the industry expanded its workforce and took a "more cautious approach" to compensation, according to a report by New York state's comptroller.
15 seconds of your voice is all that Open AIās new voice cloning model needs in order to create a synthetic, reusable version of your voice. The technology has been in development since late 2022, and is currently being rolled out to a select pool of users in the form of a text-to-voice tool called āVoice Engine.ā
30,000 in 3 months is how many times a fabricated software package was downloaded by developers. AI researcher, Bar Lanyado, noticed that generative AI has a tendency to invent non-existent software packages and decided to test what would happen if he created one of these, but with the caveat that he also injected a non-harmful malware into it. His experiment found its way into the GitHub repositories of several companies, including ecommerce giant, Alibaba.
āØļø Code Conundrum
Sponsored by Optimole, the best image optimization tool on the internet.
How good is your Python game? Take a look at the code snippet below and send me an email if you can spot the error. Please put āCode Conundrumā in the subject line. The first person who emails me with the correct answer (either saying what the error is or sending back the snippet error-free) will get a shoutout in the next issue.
Interested in sponsoring this section in a future issue? Reach out to me with āBizarro Sponsorshipā in the subject line.
āļø Tools and Resources
RunJS: A versatile desktop application for macOS, Windows, and Linux that provides a dynamic environment for exploring and experimenting with JavaScript and TypeScript. It offers instant feedback as you code, with live results shown per line, making it easier to understand what your code does. With access to Node.js, Browser APIs, Babel, TypeScript, and npm packages, RunJS caters to both learning and professional development needs. Its features, such as autocomplete, snippets, and type checking, enhance productivity, making it a valuable tool for developers seeking to quickly test ideas or debug code. It comes in both free and premium formats.
BlockNote: An open-source, block-based rich text editor designed for seamless integration into applications, offering both a robust out-of-the-box experience and extensive customizability. It supports real-time collaboration, customizable blocks, schemas, and plugins, with first-class TypeScript support and theming options. Designed to simplify the development process, BlockNote leverages the Prosemirror framework to provide a powerful yet user-friendly editing experience, making it ideal for developers looking to incorporate a sophisticated text editor into their projects.
š¼ļø Image of The Month
We previously reported on Klein Visionās Air Car in issue 135 of our newsletter, and now itās back in the news again. The Slovakia-based developer of the world's first certified flying car has just granted a production license to Hebei Jianxin, a prominent Chinese company. This will give Hebei Jianxin exclusive rights to manufacture and distribute certified flying cars utilizing KleinVision's technology within a specific geographical region. So donāt be surprised if within the next few years you start seeing flying cars roaming the skies in China.
š¬ Word of The Month
āSamfundssindā is a Danish word that literally translates into "society mind." It's the idea of putting the needs of society above your own. Given the stories we covered in this issue, especially the one about AI and the future of warfare, itās an appropriate choice. All of us who work in tech, and particularly those who are working on emerging advanced technologies, should keep āSamfundssindā in the back of our minds. For what good will our work be, if we are not helping the societies we live in?
š Results of Last Monthās Poll
Judging by the results of the polls above, it looks like the overwhelming majority of us have recognized that AI-generated SPAM is polluting the internet. Only 8% of respondents said that they basically donāt care. (Is that you, Jacky Chou? Do you read Bizarro Devs? š¤)
Well, if itās any consolation, judging by the fifth story in the top ten links of the month section, it does appear that at least some governments are attempting to take action on AI. How much of that will extend to actual AI-generated SPAM is tough to tell. My guess is that theyāll probably be more concerned with other AI-related issues and will leave the SPAM filtering to Google, but weāll see.
š§š»āš»šØš½āš»š©š¼āš» Tell a Friend or Colleague
If youād like to help us grow, just use the refer a friend button below and if any of your referrals subscribe to Bizarro Devs, then weāll not only be forever grateful, but weāll also recognize your efforts in the following way:
If two of your referrals subscribe then weāll give you a shoutout in the conclusion of the following monthās issue.
If twenty of your referrals subscribe then weāll share an article or story of yours in our ten must-see links of the month section*.
If two hundred of your referrals subscribe then weāll share an article or story of yours in our from the newsroom (top 3 stories) section*.
* The story or article needs to be relevant to the subject material we typically cover and is subject to editorial approval. In other words, it should be something that we would include in the newsletter naturally.
Thanks for reading! If this is the first time you're here at Bizarro Devs and you enjoyed the content, then sign up here to join the fun for the rest of 2024 and beyond. As mentioned in the intro, next month weāll also be bringing you a double dose of Bizarro Devs so get ready for that.
Until we see each other again,
ā Martin
Hi Martin, have been a longtime subscriber to the newsletter since before the substack days. I believe I have won a reward for referrals. How do I get in touch? Cheers, Phil