Jump In

Consider this your permission to try this or that thing you’ve told yourself, “only if I had x or y”. I know it’s intimidating to see the strings of videos that say LOOK AT THE 8 INSANE THINGS YOU CAN NOW DO WITH [insert AI model] while feeling like a luddite. To most of us, a json is a neighborhood kid back home who always picked his boogers on the bus. Forget the time and energy to go through a structured course to learn what it really means or does.

Viewed this way, it’s certainly formidable in getting anywhere. Instead, reframe it by writing down the burning questions, to interests, that have sat as embers in the back of your mind and start there. Typically, a search in any of these directions is self-directive and increasingly difficult given the SEO bloat begging to serve over giving an answer at the top of the page.

“How do I make an omelet?” you type into the search bar:

There is no direct answer. There are thousands of pages telling you they have it. It is not their fault. Search engines are retrievers. They fetch sticks. The ones covered in peanut butter come back first. It takes intelligence to find any set of options to digest, then reduce it down to answer the question more directly. Everyone from Aristotle to Einstein to Jobs to Musk and Munger have implored as much. K.I.S.S. Keep it simple, stupid.

Artificial Intelligence is a great step to address this void in transforming results. It is less up to the task in answering something with little context or needing the shotgun spread of results we’re accustomed to. Typing “Men’s shirt” into any LLM will provide a different response. Any higher order organic intelligence would respond, what do you mean?

In this sense, we’re back to trudging through nonsense for results when inserting the same thing into the web browser pulls up a whole laundry list based on history and location. This distinction is the first step in figuring how to utilize this technology, no matter your skill level, beyond basic functioning of computers.

If you want prioritized results - web

If you want more specific results - ai

There are incentives and hierarchies that are bound to influence the answers we get with LLMs that can be quite scary to think about, but most matters of truth exist on how simple they are to perceive or replicate. So, going back to our omelet.

When I ask my question again, “how do I make an omelet?”

The first thing to notice is that each answer is straightforward, and although there may be steps, a basic omelet is a basic omelet. That’s what was asked.

For those hesitant until now, this should be a great entry point as the first bit of value is getting answers versus finding your own, knowing better the distinction between the type of results you’re after. If you’d rather know what materials are best suited for colder temperatures, what brands use them, and what availability or where are they sold near me, then an LLM can help narrow.

It is the multiple arguments and follow ups that set an LLM apart. To find the same using a web browser, it becomes a mental model of searching separately, then going back for a more specific inquiry knowing one’s options. This is time consuming and often unrewarding.

Already knowing what you want up front is a helpful skill to build overall, but it’s needed to get the best out of an interaction with an LLM. The basic request for an omelet gets a basic answer. The truth is, you want a “restaurant quality” or “fluffy with sweet and savory flavors” that fit your palette, fridge inventory, and skill level. The good news is your interactions will sharpen them if you truly intend to get anywhere.

Too few realize how vague or abstract their requests are perceived. It’s easy to keep up a cycle of rationalization when two humans interact. It’s another with something that doesn’t care. A personal assistant should have known to include a pony in “find me a pretty card” while the digital one has no prior knowledge about you or care. The robots will have both types as they grow up with us. I digress…

Many will face this barrier and blame the technology for its stupidity, leaving it alone. For those left with the itch. A wish to start. Wanting to use their brain, ok with the inconveniences of learning new things, what awaits you is extraordinary.

It is less an interaction with some other presence as it is being able to tap into questions already raised and answered in the vast repositories of knowledge made by humanity. Stack overflow is riddled with poor developers hail marys answered by other kind souls. Yahoo! Answers and Reddit and Wikipedia are similar community based repositories. Then there’s the internet wide.

Instead of doing your own filtering through any number of resource(s), an LLM can do the work by providing direct and actionable requests. Then following up on these conversations so the context of the idea and project can build.

I suggest having a playground or two type thread, then keeping others for specific projects that stay on task. For those more savvy, there are an increasing number of studios, widgets, and IDEs that empower further. This does not mean the requests must be all digital in nature.

Plan a big trip. Learn how to grow tomatoes (for real this time). Finally build that thingamajiggy.

Jump in.