Intro
AI Slop 🫠, AI Vulnerabilities 😵, and AI Bubbles 😪 - All things AI 🤖
This month I didn't get to write anything for my blog because I have been working on the capstone project for my final class in completing MS in Cybersecurity 👨🎓. I am excited to be done with that soon, and I am hoping that I will be able to share that over on my blog - all of the learning and findings of my final coursework 📝.
In the meantime, I have been following the AI talk from different sources on YouTube ▶️. I think there are some curious things happening in the industry. We should be paying attention, seeing reason, watching what people are doing rather than saying, and not be led by emotions of hype or despair. Many of the things that have been happening lately do not correlate with what the industry has been promising us. We should make note of that and approach this discussion with critical thinking 🤔.
I hope you will find this interesting and educational 🙌.
Take care!
Videos
AI Was Supposed to Cure Cancer - We Got This Instead - YouTube
Vanessa Wingårdh brings an attention to an interesting phenomenon - AI companies promising amazing leaps in progress and solving humanity's greatest problems while delivering more ways to create slop on the Internet. This begs the question: "Is it due to limited capabilities and the fact that these companies are all talk but nothing to deliver on that front or is it more a matter of the demand of the market?" Maybe the slop is what makes the most money right now.
AI Coding Sucks - YouTube
With his extensive experience, CJ's brings a great perspective on AI coding as a seasoned developer himself. One of the things that stuck out to me was the fact that a lot of times the development cycle with AI doesn't yield that much benefit (especially in large codebases) because the code reviews and re-prompting often take as long if not longer than actually sitting down and churning code out yourself. The SDLC is not the same as it used to be, and in many ways we have regressed by adopting this approach.
This isn't the future I want - YouTube
One of the recent cybersecurity discoveries takes us to a fleet of poorly configured robots that operate on very poorly secured stack, which led to researchers finding some cautious vulnerabilities that should not be where something like a robot that can interact with a world around it could be compromised, and then used to compromise other robots that are nearby.
ChatGPT's Altas Browser is a Security Nightmare - YouTube
ChatGPT has recently entered the market of yet another product - Internet browser. This time they used a Chromium fork and enabled ChatGPT to access all of the features of the browsing experience. Sounds amazing, right? One thing that the researchers at Brave browser have been pointing out is the fact that this opens a door to so many prompt injection vulnerabilities where the LLM tries to interpret what it sees, but what it sees turns out to be a prompt injection asking ChatGPT to navigate to your email and send your data to a malicious actor. This technology would require many and many more iterations before it will be secure.
Breaking The Creepy AI in Police Cameras - YouTube
I have been following Benn Jordan for a while now. He is an amazing tinkerer and an independent researcher. In this video he explores ways to disable the AI that is used in OCR technologies in survelliance cameras from being able to read the plate on his car. I think this is an interesting direction in a crossover between AI, privacy, and cybersecurity.
The Attention Economy Is Everywhere. Self-Hosting Is the Escape. - YouTube
I have been interested in the topic of self-hosting different services so that I reduce the dependence on some of the common services and also get to learn how to host, configure, and maintain such services for myself in practice. One angle that I have not considered in the past was the angle of trying to reduce the attention dependence on the commonly used products such as YouTube, Instagram, Facebook, etc. Using the approach suggested by the author of the video we can start reclaiming our attention, reducing the impact of products that are designed for retention a.k.a. to be addictive, and learn how to self-host projects along the way.
The State of the AI Industry is Freaking Me Out - YouTube
In this video, Hank Green goes on about the AI bubble topic with a reference towards a graph representing the inside deals between the giants in the AI space. This was the first video that got me thinking about the implications of such a bubble bursting.
the state of the ai bubble - YouTube
In this video, CoffeeZilla dives into looking at the details of the deals between the AI giants and seeing if AI bubble is a thing or not.
Tech Billionaires Know the AI Bubble Will Burst (They're Already Building Bunkers) - YouTube
Vanessa takes another interesting look at the walk rather than the talk that comes from the tech giant CEOs and people in power in those companies. The trend that she observes in this video indicates that some of the actions of the rich and powerful suggest that they are taking financial and physical precautions in the event if something like a bubble bursting and economy crashing would happen. As always we should look where the money is rather than what they tell us to be looking at.