Morfternight #98: Incremental progress, evolution, and revolution.
The one that starts the third year.
š«š· La version franƧaise de Morfternight #98 est disponible ici š«š·
š tl;dr
Today, we get back on track. We celebrate many nice things, look at a white rose, reflect on whether AI will destroy or save humanity, and finally, on replacing binary alternatives with ternary progressions.
š¤© Welcome to the 838 new Morfternighters who joined us this month.
I love having you here and hope youāll enjoy reading Morfternight.
Share with your friends by clicking this button.
š Good Morfternight!
Well, itās been a whileā¦ precisely a month.
During this month, a few things happened. First, we closed a poll in which French won by a hair against Italian as the language to translate Morfternight. Now, I need to figure out how this will work precisely so you can prepare for some iterations and requests for feedback.
We also passed 5,000 Morfternighters. FIVE THOUSAND!
We celebrated two years of Morfternight!
I could dwell on the fact that there are 104 weeks in two years, and this being Morfternight #98 instead of #105, I missed seven. Instead, Iāll celebrate the fact I sent 97 out of 104, a whopping 93.3%
š· Photo of the week
I havenāt taken many photos lately, but thereās a good reason. Taking pictures while walking with someone else is a tricky business. We photographers often stop, sometimes backtrack, or stand there for an eternity. Detours? Oh,Ā there are plenty.
Photography is a solo gig, often at odds with other activities. These days, I walk in the company of E. or P. As much as I love making photos, I appreciate the company even more.
I often carry my camera, but itās common for me to return home without using it. A few weeks ago, as we walked through the Ferstel Passage in the first district of Vienna, a rose caught a ray of morning light in the shadows, and I was happy withĀ theĀ result.
šŗļø Three places to visit
Not the most recent articles today. I selected those a few weeks ago, but I still want to share them to continue exploring whether AI represents an existential threat or the next step of our evolution.
ā Letās keep digging into the complex world of AIās pros and cons. Today, Iāve got two relevant articles for you from The Guardian. One talks about how AI might ruin us: āFive ways AI might destroy the world.ā The other paints a brighter picture: āFive ways AI could improve the world.ā
Steve Rose interviews top experts in both articles. Whatās intriguing? Three experts show up in both pieces, proving just how complicated forming an opinion onĀ AIĀ canĀ be.
ā” Letās momentarily step away from doomsday scenarios about AI wiping out humanity. Instead, letās focus on another idea: AI might kill the web. James Vincent argues in his Verge article, āAI is killing the old web, and the new web struggles to be born.ā
I find myself nodding in agreement with the first part of Vincentās claim. Just as the web displaced magazines and most newspapers, AI is poised to overhaul the digital landscape weāre familiar with.
I canāt help but feel a sense of excitement about what weāll create toĀ replaceĀ it.
ā¢ Meanwhile, thereās some intriguing research out there on prompt optimization. One study found that simply telling an AI model to ātake a deep breathā led to significantly higher math scores.
While surprising, is it really that unexpected?
Iāve noticedāand maybe you have, tooāthat being polite with language models like GPT-4 yields better results. Itās not so much that the AI responds to politeness as these prompts provide specific context, steering the model in a better direction.
š¤ Incremental progress, evolution, and revolution.
We often face a seemingly binary choice: focus on the short or long term.
This problem appears across various aspects of lifeābe it in building a product (ship something fast or something good?), managing personal finance (enjoy money now or later?), making business decisions (prioritize short-term gains or long-term value?), or even in health choices (enjoy life now or later?).
F.Ā ScottĀ Fitzgerald famously said that the test of a first-rate intelligence is the ability to hold two opposed ideas in mind at the same time and still retain the ability to function.
I do not believe these are exclusive choices. Iāll further suggest they can be harmonized into a three-step progression.
A great example comes from the process of building a sales pipeline. The largest opportunities often take a significant amount of time to mature due to the scope of the projects, the size of the companies involved, and the substantial amounts at stake. However, awaiting these to materialize isnāt a viable strategy as it could stifle a companyās cash flow and operational momentum.
A balanced sales pipeline, on the other hand, encompasses short, medium, and long-term prospects. The short-term prospects generate a consistent influx of smaller sales; medium-term prospects contribute a regular stream of moderate-sized deals, all building the runway to close the larger, long-term contracts.
This rings true in software development as well. The most successful teams master the balancing act of maintaining their existing codebase to offer a flawless user experience, iterating on the current feature set to stay relevant while simultaneously dedicating resources to deeper, groundbreaking innovations that elevate them ahead of the competition.
New technologies are not exempt from this pattern.
Instead of deliberating whether the AI revolution is already here or far beyond the horizon, I prefer to view AI as a set of tools. These tools can reliably enhance my efficiency at the margins immediately, amplifying my output in my areas of expertise.
While I canāt fully trust them in less familiar fieldsāthey sometimes fabricate alternative realitiesāI am optimistic. As they improve and I become more adept at utilizing them, I anticipate that, in the medium term, theyāll enable me to tackle a broad range of tasks comfortably. In the long term, there is a possibility they get so efficient that I wonāt be needed anymore, which doesnāt mean Iāll be dismissed. I still know a few good jokesā¦