How Is Machine Learning Different from AI and Algorithms?

Weekly Wheaties #2602

Jan 11, 2026

In this newsletter:

📝 Post: How Is Machine Learning Different from AI and Algorithms?
🗞️ In Case You Missed It: CES
🗞️ In Case You Missed It: ryan3000
🗞️ In Case You Missed It: Headlines
😎 Pick of the Week: Sports, Games, and Hobbies
📦 Featured Product: Belkin Stage Powergrip

📝 How Is Machine Learning Different from AI and Algorithms?

As often as I talk about AI, there are two other areas of this space that need discussing that may help show the bigger picture. A common term in this space is LLMs (or Large Language Models), but I would first like to discuss another term - Machine Learning.

Machine Learning has been around for much longer, and was a term used by the larger tech companies (e.g., Apple and Google) for years, but the term was a bit too broad for consumers to really understand it as the need wasn’t there.

As AI has become more mainstream and consumer-friendly, I think it will be helpful to understand Machine Learning. In the past, I’ve briefly discussed how Google works by way of the democratic rule. In short, the more people who search and ultimately visit a specific website from their search result last, the more that site moves up the results in popularity. The next time someone in that similar demographic area searches the same term, the site’s value from the first person causes it to rise for the next. And so on…

This is also how Google and other search engines know if you’re looking for an apple fruit or the Apple company when typing in the word “apple.” Your previous search history helps here, too, along with other demographics your account and computer tell about you.

Another term in the same concept you have heard of is an Algorithm. An algorithm is a very specific set of rules that the computer must follow. In the search example, every time a site is visited, it may get another tally mark. And the algorithm adds up the tally marks from others, looks at other information, combines those tally marks, and boom - here’s your search results!

What’s actually happening with Machine Learning is the computers are ‘allowed’ to create their own rules. Rather than a programmer giving computers a set of rules (like an algorithm), they are given tons of examples and data. This allows them to spot patterns and help them make predictions. Just like humans learn better from practice, a computer can, too.

Spam filters work similarly. They look at the email name, address, titles, and what’s included in the body of the email - and especially things in the email humans can’t see (the “code” of the email). They are also looking at where the emails are sent from, and how many other people may receive similar emails from that same person/area/context/etc. However, that isn’t the whole story. At least from the consumer side of things.

There are literally hundreds of billions of emails sent every day. And there is no one main “spam filter” company. Microsoft’s Outlook filters work differently from Google’s Gmail filters, and you may not like it or believe me, but the paid versions are different than the free versions. Even the paid personal plans are different from the paid enterprise plans. But why does that matter?

Essentially, a spam filter is not a “percentage” out of 100 on how well it works. No spam filter will be 100% accurate. Instead, these spam filters (and machine learning by context here) typically may have a scale appropriated to it. Instead, what is happening is these machines are prioritizing a few different things. In the email spam filter case, it may be strictness versus leniency, safety and security, the risk of blocking legitimate messages, and the overall user experience.

All of these things combined allow the system to either let more emails in (perhaps also allowing more spam in accidentally), or keep as much out as possible (possibly blocking real emails by accident). As the machine learns, it gets better.

Some places you may already be using Machine Learning include: recommendations from your streaming entertainment platforms, search results, and navigation apps, among others.

There is technically a lot of math happening on the backend with Machine Learning, but it’s not quite as much as AI. Machine Learning is also not creating something new like AI, and it may sometimes come across as blatantly wrong due to the differences in results from a Machine Learning computer versus how an algorithm works - the same way two navigation apps may show different paths to the same location, or the same search term has completely different results from Bing to Google. Patterns (like Machine Learning use) do not always follow rules, especially when humans are involved in regard to social context.

Knowing when or why one is used over the other can be pretty obvious. Algorithms follow rules. Machine Learning learns patterns. A basic example is to create a rule (algorithm) in our email to block any incoming email with “You win” in the subject line as spam. However, letting Machine Learning take over may block incoming emails with “You’ve won” and also be marked as spam. As you can see, though, neither is perfect.

So when technology gets something wrong, the real question isn’t “Why did it fail?” It’s, “Was it following rules, or learning patterns?”

🗞️ ICYMI: CES

The 2026 Consumer Electronics Show (CES) was held last week, announcing a ton of futuristic and AI-enabled products. You can check out CNET’s post on Everything Announced From Best of Show to AI Toys and Dancing Robots, The Verge’s video recap of The Best of CES, or TechCrunch’s most bizarre tech announced so far at CES 2026.

To highlight a few cool new things coming this year:

Lego announced the Lego Smart Play System. This is composed of Bricks, Tags, and Minifigures. A single 2x4 brick houses Bluetooth, an NFC tag, a microphone, a speaker, lights, and more. If you are in the Lego space, check out Rebrickable and check out possible alternate builds for the sets you already own.

Clicks released new MagSafe Keyboards and the Communicator, a text-first smartphone that runs Android. Check them out on YouTube!

Pickle announced their Pickle OS and Glasses and dubbed them “The Era of Soul Computing.” These are some fun new smart glasses with a fun name to match. Coming later this year.

🗞️ ICYMI: ryan3000

One of your favorite family-friendly YouTubers, Ryan Trahan, started a new channel - ryan3000 - where he vouches to spend 20 minutes a day in his Minecraft world, every day in 2026, in hopes of ending the year with the same number of digital chickens as he has subscribers. He’s had his first recording issue over the weekend while trying to record on the road, but I have hope. He’s almost at 100,000 subscribers, so if you’re a fan, go subscribe!

🗞️ ICYMI: Headlines

Streaming

😎 POTW: Sports, Games, and Hobbies

Looking to start the new year enjoying a new sport or hobby? You could start by browsing reddit, find a game from this list on How To Win Classic Family Board Games, Go for the Unclaimed Titles of a few Guinness World Records, or check out The 100 Best Sports Moments of the Quarter Century.

If you want to learn Chess, I suggest either Chess.com or Chessfish.io. Don’t forget your Magnetic Chess Board, or the Rolls Royce Chess Set if you have some leftover Christmas money.

Another favorite of our family is Sequence, but may I suggest the Jax Giant Sequence Board Game.

📦 Featured Product

If you are one of those people who buy new phones mainly because of the upgraded cameras, or you choose to use your mobile phone as your main camera, this is for you! The Belkin Stage Powergrip is more than just a camera grip for your phone. It also functions as a wireless charger for your phone, can charge other devices via USB-C, functions as a vertical stand, connects via MagSafe, and can also mount to a tripod for even more stability. This Magnetic Phone Camera Grip Handle is a less expensive option that doesn’t include the battery backup, but does come with a brighter LED light with two color temperatures to choose from. Both connect over Bluetooth in order to control your camera’s shutter.

Discussion about this post

Ready for more?