View profile

The Monday 6 - September 6, 2021

Kyle Petzinger
Kyle Petzinger
Happy Labor Day, US and Canada-based friends!
I hope you enjoy your restful, long weekend.
Onto the 6:

1. McDonald's Ice Cream Machine Update
If you’ll recall, back in May (#1), I shared a video exploring why McDonald’s ice cream machines are always broken. If you didn’t watch the video then, check it out now. It’s well done and quite entertaining:
The REAL Reason McDonalds Ice Cream Machines Are Always Broken - Johnny Harris - YouTube
The REAL Reason McDonalds Ice Cream Machines Are Always Broken - Johnny Harris - YouTube
As a quick summary, essentially, McDonald’s and the maker of the very particular model of ice cream machine used in every store in the county have an exclusive deal with one another. This contract (conveniently) requires technicians from the ice cream machine manufacturer to come out and fix the machine when it “breaks”. The clear conflict of interest and potential for abuse has led to the scourge of broken McDonald’s ice cream machines around the country.
But as a follow-up from that story, the Federal Trade Commission is nowinvestigating the matter, the first sign the core of the issue may be solved! Let’s hope any potential antitrust suit brought is more substantial and less confused than the FTC’s suit against Facebook refiled in August.
McDonald’s Ice Cream Machine Is Broken (Again). Now the FTC Is On It.
2. Zoom And Enhance, But In Real Life
Remember watching CSI, circa 2004, and watching this nonsense magic?
Zoom...enhance.
Zoom...enhance.
Well, Google’s AI team is basically doing this in real life. The specialized AI is fed a low-resolution photo (as low as 32px X 32 px) and interpolating up to images large enough for actual use.
Example of the cascading pipeline that includes a sequence of diffusion models: the first generates a low resolution image, and the rest perform upsampling to the final high resolution image. Here the pipeline is for class-conditional ImageNet generation, which begins with a class-conditional diffusion model at 32x32 resolution, followed by 2x and 4x class-conditional super-resolution using SR3.
Example of the cascading pipeline that includes a sequence of diffusion models: the first generates a low resolution image, and the rest perform upsampling to the final high resolution image. Here the pipeline is for class-conditional ImageNet generation, which begins with a class-conditional diffusion model at 32x32 resolution, followed by 2x and 4x class-conditional super-resolution using SR3.
Check on the video on their blog post to see more examples and explanation:
High Fidelity Image Generation Using Diffusion Models
High Fidelity Image Generation Using Diffusion Models
Google's New AI Photo Upscaling Tech is Jaw-Dropping | PetaPixel
3. JFK Assassination Conspiracy Theory
I’m a bit young to ever get that far down the JFK assassination conspiracy theory rabbit hole. Still, evidently, this photo has been used as a main exhibit in the case that Lee Harvey Oswald wasn’t the actual killer:
Skeptics believe this photo shows obvious signs of manual manipulation, including irregularities in the shadows, Oswald’s face, and more.
Check out this Vox Darkroom episode on YouTube to learn more and see why ultimately (as with virtually every conspiracy theory), these theories don’t hold much water.
Why people think this photo of JFK's killer is fake
Why people think this photo of JFK's killer is fake
4. Pictures Of Super Magnified Things
Look at a bunch of everyday items super magnified. Among others, you’ll see the following items ultra up close:
  • Coffee
  • Salt & pepper
  • Sharkskin
  • Mascara brush
  • & more!
Just look at this shot of a snowflake!
 How do they actually look like this?
How do they actually look like this?
Weird Photos Of Stuff Super Magnified
5. Find A Movie You'll Both Love
I love a good single-use website like this. Date Night Movies prompts you to enter two movies: 1 you like, and 1 your partner likes.
The site’s algorithm will then combine the picks and find a movie that will satisfy you both.
Movie Recommendations for Date Night
6. Practical Career Advice Twitter Thread
Check out this Tweet thread of 20 principles & axioms to keep in mind as you’re building your career.
Romeen Sheth
Early career years are painful.

You feel like an idiot 98% of the time - lost, confused and insecure.

I wish I had a cheat sheet of principles for my first job.

So I put one together.

Here are 20 things about building a career I wish I knew sooner:
#13 was the most resonant for me:
Romeen Sheth
Principle #13: The answer is just as important as the process

Nobody likes to work with a “know it all.”

Don’t compromise on the solution - the best answer is key to driving impact.

But getting to the right answer without any buy in or commitment is worthless.
Give the whole thread a read. I think you’ll find something useful in there.
That’s it for this week. If you found anything useful or fun, let me know! And if you have any suggestions, shoot them my way. 😊 See you next week!
Did you enjoy this issue? Yes No
Kyle Petzinger
Kyle Petzinger @kylepetz

Every Monday, you'll receive the 6 articles, podcasts, tools, products, and more that I found interesting that past week.

If you don't want these updates anymore, please unsubscribe here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Created with Revue by Twitter.