Unfog Your Brain: 7 Best Tips to Fight Clutter (Garbage-In Garbage-Out)

With the rising interest in generative AI, all of us are seeking signal in a noisy world.

Those passionate about AI understand that a model’s effectiveness is only as good as its input training data. The principle of “Garbage in, Garbage out” applies here.

If you regularly consume superficial news with a negativity bias, it’s easy to predict how your brain might respond to new information: with rejection, doubt, and criticism, often stemming from impulsive thinking.

I receive statements like these: “… I read that AI already becomes dumber, and it only averages all available information without producing anything new or creative …”

If you point people to research indicating that ChatGPT ranks higher on creativity tests, they argue that the tests must be faulty. Yeah, sure they are faulty. Still, it’s better to go with Occam’s Razor, embrace the simple explanation, and remain open-minded in the face of change.

It may sound amusing, but many people dismiss the idea that Large Language Models (LLMs) are a disruptive force, set to multiply their impact tenfold every two years, thanks to the exponential decline of AI training costs.

The contrary to “Garbage-In Garbage-Out” also holds: if you train a model (or your brain for that matter) on high-quality data, you’re much more likely to produce high-quality outputs.

Quality-In Quality-Out

So how can you cleanse your training data and, thereby, boost the performance of the AI model that is your brain?

Here are some strategies I employ:

1. Go for long walks. After a walk in nature, I always find clarity about my next step. Taking one step at a time, with a distant long-term goal in mind, simplifies life. Clarity becomes increasingly vital as the influx of low-quality information leads to confusion, brain fog, and fatigue.

2. Think slowly and take time. Many people form shallow and hasty conclusions, presenting them with undue confidence. I’ve learned that such quick opinions, though easy to offer, often hold little value.

3. Listen to podcasts with field experts but only on topics within their expertise; I avoid those who speak outside their domain. Don’t necessarily listen on Bitcoiner’s advice on food or the money tips from broke people (you may be able to leverage the powerful technique of inversion though: do the opposite).

4. Occasionally browse Twitter/X – but not too often, as it can be a time sink. Among the noise, you might find a valuable insight, but I’m cautious with social media. Better yet – avoid all social media including X.

5. Read select tech and coding newsletters. I strive to offer such a newsletter with a high signal-to-noise ratio – you be the judge. Join us free!

6. Skim ArXiv and new research papers: these are often a rich source of knowledge and remain some of the highest-quality information available.

7. HackerNews is also an excellent source for straightforward news with a high signal-to-noise ratio.

This brings us to today’s Finxter tutorial if you’re interested in building your own HackerNews-like site:

πŸ”— [Blog] How I Coded a Hacker News Clone in Django

You can check out our email newsletter with content like this for free by downloading our Python and ChatGPT cheat sheets here: πŸ‘‡