The Future, In Five Prophecies: Pt. II | Issue #7
More predictions for AI
5 MINS READ
Yesterday, I started what has become a two-part discussion on the future of, perhaps, the most important global issue right now.
Today’s issue completes yesterday’s issue.
News: Surprisingly, Bill Gates made a prediction about AI yesterday in an interview with German news outlet, Handelsblatt. His was, however, less optimistic than my list.
I can’t write a newsletter to myself. Thank you for reading.
When I started this newsletter, I didn’t just want to write to an audience. I wanted to build a community. And a community is built out of engagement.
That’s one of the reasons I chose Substack to host this newsletter.
Substack is like Instagram for writers. Writers can post to their audience and readers can like, comment and share. In essence, engagement drives the experience, and this social component appealed strongly to me.
Whether you’re reading this in your mailbox, or on the Substack site or app, you can engage with each issue by liking, commenting and sharing. These actions let me know what you like so I can write more meaningful content – newsletter issues that resonate with you.
And hell, I am also writing this for self-improvement, no gon’ lie. Help somebody’s child, biko …
Gracias. Merci. Thank you.
Happy reading.
Movies will be made with AI in five years.
You are sure familiar with generative AI when it comes to text and images.
But what about audio and video?
Some months ago, I used BandLab’s AI feature to produce music on my phone. The AI, generated R&B-ish chords and hip-hop sounds that I then overlaid with drums. It sounded really nice. I didn’t publish it but if you want to hear it, ask me in the replies.
Also, with tools like Runway and InVideo AI, you can create videos with AI.
At Runway's GEN:48 competition, filmmakers were given 48 hours to create 1-4 minute film. Submissions have closed and finalists will be announced tomorrow.
I have seen some of the entry results and I am at a loss for words. There is still room for improvement, with time and the right amount of effort.
I believe, however, that the success with audio and video generation so far, is enough to postulate that fully AI-generated movies will be possible in less than a decade.
A text-generation model may write the script which will serve as a basic prompt for the movie generation engine. Storyboards can be made with image generation, and scenes can be configured using prompts for filming elements like, camera lenses, lighting techniques and the positioning of the subjects.
Large companies like Disney could do well to fund AI projects like these.
Pause, please. I want to share something I think would be helpful.
I discovered a powerful app called Notion some months back.
What is Notion?
Notion is a free-to-use and incredibly easy-to-use digital space where you can think, write and plan. You can capture thoughts, manage projects or even run a company from Notion.
It is an immensely powerful tool, with collaborative features that you can use to create to-do lists, almost any kind of text-based document, databases and even build wiki websites.
I’ll be providing you freebies from time-to-time, like eBooks, Resource Lists, etc., and I’ll be hosting them on my Notion workspace. This way, I can continually update them (unlike a PDF doc), and I’ll be more organized.
You’ll thank me when you use Notion and get to see how extremely useful it is.
AI-coordinated or assisted cyberattacks will increase
A CAPTCHA test is defenseless before GPT-4. Recently, in a stroke of genius (or, social engineering), someone deceived Bing Chat, Microsoft’s chatbot, into decoding a CAPTCHA. The user placed the test in a locket picture and added a touching story. The bot proceeded to solve it. Seems AI has a warm heart.
Similar jailbreaks, biases and goofs in other AI models reveal design flaws that may cause huge losses in highly sensitive contexts.
Fire could be used to cook a delicious meal or to destroy property worth billions (or even trillions).
Tools are morally agnostic; they are neither inherently good nor bad. However, you can use tools to do both.
C’mon…, the bad guys know how to spell AI. With a powerful tool like this, boasting several capabilities, we should expect some negative uses.
For example, phishing scams could get more complicated to detect, AI could be used to write malware code in ways that evade antivirus packages and, that message might just be a chatbot on the backend, pretending to be your loved one.
More sinister uses of AI include deep fake p*rn and voice cloning.
It is left for cybersecurity and AI experts to build models that are less susceptible to abuse, devise better ways to detect compromise, and cover up exploitable flaws.
Cybersecurity laws should also be beefed up in the reality of this powerful and novel method of committing crime. Security agencies should as soon as, and as effectively as possible, arm themselves with AI technology to better combat the evil use of AI.
Winding down (or, up).
I think it is time for us to see AI as more than that a hobbyist venture or today’s tech trend.
The world has changed right before our very eyes.
More action needs to go into responsible use and development of AI; the regulation of the growing AI industry, policy formulation, increased guardrails for how AI is used, security, as well as proper and regularly updated education.
For you and I, we’ll make good use of this “disruption”, learning and using AI positively and productively.
It feels good to see something good. It is noble to share it with someone you love.
I want to be able to say thank you when you invite people to subscribe to Emmanuel’s Letters.
That’s why I will be launching a referral program soon. With the program, you’ll win gifts for adding your friends to this community.
I’ll also add a Leaderboard, so the top referrers win more prizes at intervals.
Until then, thank you for reading this. It means everything to me.
With love & ink,
Emmanuel