Today, let’s veer away from the world-altering article and models to discuss something equally pervasive but often overlooked: recommendation systems. Although these systems are omnipresent across the websites I frequent—suggesting what I "might like" and filtering out what I "might not"—their varying algorithms can lead to some amusing paradoxes. For instance, I have a distinct preference for crisp apples over mushy ones, but it’s unlikely that any recommendation system could discern such subtle distinctions. And then, there’s the almost comical scenario where after purchasing a refrigerator, e-commerce platforms continue to bombard me with deals on more refrigerators.

However, what I’m keen to delve into today isn’t just the flaws inherent in these systems. I recently came across an article by a neuroscientist discussing how to monetize our dopamine levels. He will never say, “We can make money off your dopamine.” Modern social platforms and video websites are adept at converting our attention into advertising revenue through mechanisms like short videos and the pull-to-refresh feature. It’s no wonder the title of this blog post is “Attention is All They Need.”

The landscape of fame has shifted dramatically. In the past, becoming a celebrity might have required a particular skill or talent. Today, it seems to be more about capturing and maintaining the audience’s fleeting attention. Quick edits, high saturation, and overt displays that cause both visual fatigue and pollution seem to serve as a sort of narcotic for our nervous systems—they’re addictive and ultimately, quite futile.

The internet once exposed me to unknown places, unfamiliar jobs, and ways of living from past eras that I had never considered. I could explore fields I’d never stepped into and acquire knowledge I was eager to understand. Now, even if you show interest in something as niche as the Fibonacci sequence, you’re more likely to be recommended a slew of short videos on how to quickly multiply 43 by 23. Clearly, the minds behind these recommendation algorithms are not fools, but the systems they’ve created are far from perfect. In this highly automated age, finding truly meaningful content often requires sifting through more noise than ever before.

Maybe it’s just the overwhelming number of cat memes recommended on the homepage of my video streaming service, but sometimes, I can't help but voice my grievances.

Until next time, peace out.

The ability of humans to create and use tools has propelled our development far beyond that of other species, establishing us as the dominants of Earth. A famous observation highlights that despite humans being far less energy-efficient in running compared to other animals, the introduction of the bicycle significantly surpassed the energy efficiency of any animal. This example profoundly illustrates the concept that tools equal efficiency: the appropriate use of tools within our limited time and energy determines the heights we can reach.

Tools come in a vast array, primarily designed around the needs of humans as the operators, emphasizing our leading role in their usage. Therefore, perception towards tools is crucial. For instance, dining habits vary across cultures—from chopsticks and spoons in some regions to forks, knives, and direct hand use in others. This variety shows that our hands are not just crucial tools for interacting with the world but also play a key role in writing this piece.

As software developers, we are faced with a plethora of tool choices, from editors like Vim, Emacs, VSCode, to programming languages like Java, Python, JavaScript, and various frameworks, algorithms, coding standards, and testing metrics. With each project completed through a series of selections, we invariably question: Are these tools truly the best choice?

I personally lean towards learning fundamental yet versatile tools to achieve as many development goals as possible, seemingly the optimal choice. Take chopsticks, for example, whose design principle is simple yet highly practical. They can perform a multitude of tasks beyond just picking up food, such as cutting fish or picking difficult-to-grab items, showcasing the ingenuity of tool design. However, even such a tool requires proper technique for effective use. This epitomizes the idea of a one-time investment for lifelong benefits, with a single pair of chopsticks catering to a wide array of East Asian dishes. Nevertheless, for certain foods, specialized utensils like a fork and steak knife are obviously better suited.

Faced with an ever-increasing array of programming languages, frameworks, and tools, the root of choice paralysis lies in our difficulty digesting such a vast amount of information. But as individuals, do we really need to master all these tools? Without them, are we incapable of developing excellent software? I believe not. Often, a computer and a suitable editor like VSCode are sufficient. What's needed can be learned as required. This approach, though possibly myopic, allowing for rapid learning and adaptation, may lead to regrets about not choosing a better architecture sooner, as I've experienced through restructuring my personal website three times. However, what did I learn from this? Are these experiences truly beneficial? These are questions I often reflect upon, yet they still hold value as mental exercises.

Let us, within the confines of our limited time and life, rapidly iterate products and continuously ascend in a spiral manner, possibly the correct path to reaching higher realms.

I recently finished binge-watching 'Person of Interest', and the entire experience was a non-stop thrill ride - the brilliant storytelling left me in awe episode after episode. The demise of certain characters really tugged at the heartstrings. This show seems to underscore that good ultimately triumphs, though the fragility of life invokes deep introspection.

Going in, I viewed this as a Batman-esque vigilante show: a badass crime fighter, a tech genius sidekick, with each episode serving up a unique case-of-the-week. But it had so much more - espionage elements, laugh-out-loud humor sprinkled in, and the power struggles between the various factions provided constant high stakes.

However, the later seasons pivoted to explore the ethics of artificial general intelligence and whether humanity should cede control to AI. One view I agreed with wholeheartedly was, "It can never have human morals because it's not human." We may be flawed and inefficient, but we're not machines. No amount of training data can bestow true empathy and moral reasoning upon an AI.

So would an AGI-controlled world actually be better? And how mind-blowing would an AI vs AI war get? This decade-old show did a remarkably prescient job examining these issues over its 103 episodes. From my vantage point, if push ever came to shove in an AGI conflict, it could make the show's depiction look tame. That said, ramping it up further may have strained credulity for audiences.

Looking at today's landscape, with big tech gobbling up immense AI compute and energy resources, trying to envision what an recursively self-improving AGI could become is mind-bending.

Based on "60 Top-Quality Singles of 2023 You Shouldn't Miss | Prism Scenery Sp.10 - Bilibili".

I've compiled a list of my favorite music tracks featured in this video. My primary interest lies in rock music. Additionally, I have a fondness for full-length albums. Below, you'll find the Spotify links for these albums. While there are certainly other albums worth exploring, I don't claim to be a comprehensive music radar. However, I'm on the lookout for more summary videos to enrich my playlist.

Siren Spine Sysex - Proc Fiskal​

The Record - Boygenius​

Trace Denim - bar Italia​

My Big Day - Bombay Bicycle Club​

Re: This Is Why - Paramore​

Burning It Feels Like​

Dis You Know that There's a Tunnel under Ocean Blvd - Lana Del Rey​

I've Seen a way - Mandy, Indiana​