The Blacklynx Brief - Vol.7 - ChatGPT losing traffic and 2 heists in Vegas

Welcome to Volume 7 of the Blacklynx Brief.

Another wild week in cybersecurity that saw the biggest casino’s in the world fall victim to a vicious ransomware attack and an even wilder week in AI where it seems OpenAI is getting pummeled with lawsuits and declining numbers. BUT at the same time they made a huge announcement just yesterday.

Cybersecurity News

Microsoft AI Researchers Accidentally Expose Data

Someone from an AI research team at Microsoft uploaded a huge amount of data into Azure/Github (a Microsoft Cloud solution). Due to a misconfiguration they accidentally exposed 37TB of data to the public. The data includes passwords, private keys, secrets and more than 30.000 internal Microsoft Teams messages. Yikes.

Vegas Casinos Hit Hard

In what seems like the cyber variant of Ocean’s Eleven both MGM Resorts and Caesars Entertainment became the victim of a fierce ransomware attack. The ransomware gang believed responsible is the infamous “Scattered Spider” - a group believed to be made up of young adults in the US and the UK. Caesars Entertainment reportedly paid millions of dollars in ransom. Both companies are expected to perform a regulatory filing at the SEC with more details on the attack . MGM Resorts is still keeping systems offline across their more than 30 hotels and casinos worldwide.

Criminals are slowly cracking Lastpass Vaults

Not the latest news but something that needs mentioning. In November 2022 password manager Lastpass disclosed a breach in which hackers stole password vaults containing both encrypted and plaintext data for 25 million users.

Since then, there has been a consistent series of six-figure cryptocurrency thefts aimed at tech-savvy individuals prioritizing security. This has led some experts to believe that criminals might have successfully accessed some stolen LastPass vaults.

Taylor Monahan is the lead product manager at MetaMask, a renowned software-based cryptocurrency wallet for the Ethereum blockchain. From late December 2022 onward, Monahan, along with other experts, has pinpointed a distinct pattern of evidence linking thefts that impacted over 150 people. Together, these victims have lost crypto assets worth over $35 million.

Greater Manchester Area Police Data Compromised

Few weeks ago we mentioned a breach of London’s Metropolitan Police. This time a cyberattack that targeted an ID-card manufacturer has exposed the personal information of Greater Manchester police (GMP) officers. The UK National Crime Agency is investigating the incident but it’s safe to say that someone is targeting these police entities. This is a very dangerous trend obviously because this also exposes undercover officers and potentially will cause life-threatening situations.

Geopolitical Nefariousness

It’s always difficult to select the content for this newsletter because there is so much going on at all times that it can really get confusing. As an illlustration here is a list of shenanigans with a geopolitical edge that was discovered this week alone:

• Pakistani threat group “Transparent Tribe” targets military and diplomatic personnel in India and Pakistan

• Beijing is laying the groundwork for mass disruption targeting the US

• China-Linked Actor Taps LInux Backdoor in Espionage Campaign

AI News

Adobe AI News

Adobe has unveiled some powerful new features in two of their products: After Effects and Premiere Pro. One of the features is called “Enhanced Speech” in which AI can remove background noise and improve the quality of poorly recorded audio. Something that before used to be done by professional sound engineers. In Premiere Pro thanks to the improved Text Editing feature you can automatically let AI edit out the “uhhms” and “aaahs” and other stop words.

Adobe Firefly has come out of beta and is being integrated into the Adobe suite of products as well as being added to the Adobe Creative Cloud suite of apps , together with Adobe Express Premium, which is basically Adobe’s response to the success of Canva. They don’t even try to hide it - the UI looks more or less the same. Adobe wouldn’t be Adobe without using these new features to hike their price rates. I’ve your on the monthly plan - your price just went up with 7,5€.

ChatGPT is losing traffic - is the hype dying down ?

According to data from analytics firm Similarweb, traffic to ChatGPT has dropped significantly in the last three months. However, in an update on top of that article there is a bounce back as the new school year resumes and milllions of students are resorting to the model for completing their homework.

Still, we’re trending down.

ChatGPT’s developer OpenAI has another lawsuit on its hands

Just like we saw earlier in the newsletter - there are cybersecurity challenges with AI but also legal challenges and implications. The OpenAI legal team has been quite busy over the last few months and the lawsuits keep piling up. This time it’s a group of writers that has filed a class-action lawsuit against OpenAI for copyright infringement.

Meta is trying to one-up OpenAI

And to top off a miserable week for OpenAI, Meta has announced the development of its own Large Language Model, Llama 2 and has already hinted it is going to be exponentially more powerful than GPT-4. It seems that what Elon Musk has predicted is coming true : because of capitalist and commercial incentives tech companies are racing to develop the very best AI model, probably taking shortcuts in the safety department. The conditions for the creation of an uncaged beast are slowly shaping up.

Dall-E 3 is coming to ChatGPT (and yes it rhymes !)

Midjourney is getting some fierce competition as OpenAI has announced on Wednesday that their own image generation engine will be integrated in ChatGPT if you’re a Plus or Enterprise user. This is a major upgrade. You’re not only able to use plugins and the code interpreter as a plus user. You now also get image generating powers.

Authors have to disclose the use of AI on Amazon Kindle

YouTube is filled with videos on how to “self-publish” a “novel” using ChatGPT. It’s a worrisome development as indeed - in search for a quick buck- the market is flooded with AI-generated text that is sold as a “book”. A tactic that actually works. As a response to the trend Amazon is introducing a new policy requiring Kindle Direct publishers to disclose in what matter they have used AI.

Consultants with ChatGPT > Consultants without ChatGPT

You’ve probably seen this graphic hit your LinkedIn feed this week. A recent report by Boston Consulting Group (BCG) in collaboration with Ethan Mollick of the "One More Thing" newsletter reveals that consultants using GPT-4 outperform their peers, achieving staggering improvements.

✅ 12% more tasks
✅ 25% faster results
✅ 40% increase in productivity

Cheating? Or skill enhancement?

Predicting disease from retina scans

Researchers have developed an AI tool called RETfound that is capable of detecting eye disease and risk of Parkinsons based on an image someone’s retina. This is another fascinating example of how AI can make our lives a lot better.

That’s it from us !

If you have learned something new this week - please pay it forward by performing a simple share on LinkedIn.

I’m not asking for your soul - just a little share :)