Happy Thursday!

For the past few years, I've heard complaints from some of my peers about how dating apps "don't work".

If anything, the complaints only intrigued me to use dating apps more. But I only started using dating apps efficiently after I started learning about machine learning and AI.

Today, I'll delve into:

1. What Neural Networks Are
2. My 4-Step Neural Network Decision-Making Framework
3. Pitfalls to Avoid

What is a Neural Network?

Our brains contain a dense network of neurons.

When one neuron activates, it fires signals to other neurons that it's connected to. These patterns of activation give rise to our brain functions, from thoughts to speech and movement.

A neural network is a computational model that's inspired by how our brain works.

It's made of at least 2 interconnected layers of neurons that can produce specific patterns of activation in response to a given input.

Using Neural Networks for Decision-Making

The 4 steps in this framework are:

1. Begin with the end in mind
2. Break it down into smaller milestones
3. Prioritize 3 - 5 criteria for the milestones
4. Use the model and adjust whenever needed

1. Begin with the end in mind

In as much detail as possible, what's the ideal outcome you want to have from making the "right" choice?

For this example, my end goal of using dating apps is to find a romantic life partner.

2️. Break it down into smaller milestones

To find a life partner (eventually), I need to:

5. Love them for who they are

4. Meet them several times in person

3. Have likeable online conversations

2. Be matched with them in the dating app

1. Filter out suitable profiles from the rest

3. Prioritize 3 - 5 criteria (neurons) for the milestones

Milestone 1 is to "Filter out suitable profiles from the rest".

My personal green flags for this stage:

✅ No weird responses

✅ Similarity in lifestyles

✅ Thoughtful responses for their profile

A criterion can be a hard yes/no or a percentage. You can also prioritize some criteria over others.

• If a profile says that we could "sleep next to a lake" together, that's an immediate No (safety first)
• Another hard No for me: Anyone who smokes or does drugs.
• Those who drink alcohol occasionally are okay. If their profile looks thoughtfully put together, I might send them a like.

Milestones 2 & 3 are not within my control, but if:

• We have some common ground
• Their energy and interest levels match mine
• Our online conversations did not raise any red flags

I'll ask to meet them in person to get to know them better (Milestone 4).

4️. Use and adjust whenever needed

As I used my neural network more and more:

• I found more criteria to consider. For example, I found that profiles with suggestive pictures were also more likely to be interested in short-term hookups, which is not aligned with why I'm using dating apps.
• I dropped some criteria. I used to consider physical appearances when filtering profiles. I don't consider that now; if their profile was put together well, I'll send them a like.
• I adjusted the weights of some criteria. If we don't have enough common ground to sustain an online conversation (e.g. similar topics of interests), it will be even harder to sustain an in-person conversation. So, I increased the threshold for the "common ground" criteria.

Remember: You're always in control of your criteria and thresholds. Adjust them to be more useful to your own life.

Pitfalls to Avoid

Avoiding intangible criteria because they're hard to define

Building a neural network for choosing job offers seems easy: Pick the one with the highest pay and work-life balance.

But a lot of intangible, invisible factors also contribute to a great work experience:

• Compassionate company leaders can make a huge difference
• Company perks lose their novelty after the first few years
• Commute can suck a lot more energy than you realize

Defining intangible criteria and their thresholds takes time and experience. But you should not avoid them—especially if they're consequential.

Giving every criterion the same weight

This slows decision-making. If you prioritize everything, you prioritize nothing.

If you want to make decisions fast, have dealbreakers and/or criteria you're willing to trade off.

Lowering your thresholds merely to "increase opportunities"

Set high standards and don't settle. Don't lower your thresholds just because someone told you you're too "picky".

Some exceptions:

• If it's a career decision and you have deadlines to meet, you can temporarily lower your thresholds. Example: When your savings are running out and you're deciding between job offers.
• If your thresholds are based on unhelpful biases—reconsider those thresholds.