26 Comments

I already regret writing this but email newsletters, like the post office, have the romance of never rescinding the letter

Expand full comment

why do you regret writing?

Expand full comment

You could always say something 10% clearer with 30 more days of thinking

Expand full comment

just keep editing and the email version works as an exclusive first-draft preview for subscribers ha

Expand full comment

You have such a wonderful, unique, perspective on things, thank you for sharing your writing publicly.

Expand full comment

I think it's helpful to treat it like an orchestra and you're the director. It pulls up lots of synonyms but that's all it does. If you need synonyms it's great but life is more than that.

Expand full comment

funny enough I actually made my own website for synonyms: https://carefulwords.com

Expand full comment

It's great. You should randomize the three words

Expand full comment

Nevermind, you hardcoded in the examples

Expand full comment

Some words have no associated quotes at all, so I wanted to make sure people saw examples with a good spread of words + quotes

Expand full comment

If you're hardcoding in that shallowly then you need to universalize your code

Expand full comment

To chase of the lossless gives itself good examples in audiophilia. Good thoughts, myself still resisting to use Ai for anything except painting or photos generation I see the truth in sometimes feeling with 6th sense sometimes just giving it time, we live in times when problem presented to us is not solvable. Future is here, only its distribution is not even.

Expand full comment

Become unsummarizable.

Expand full comment

To the extent Tim Urban spent ten minutes to see if it was worth spending an hour watching the full video it's a reasonable bet. He did a follow up tweet (https://x.com/waitbutwhy/status/1873583821217746957) which seems to support my hypothesis: "To the commenters disappointed in me for not watching the whole video: This was not a core piece of my research. While writing, I wanted clarification on something I had already researched. A searched turned up a video that seemed relevant and I wanted to know its basic points!" I agree with the risks you outline and the value of sweating the details, but AI summaries offer another "point on a trade-off curve" that may be worth using from time to time.

I think there are a broader set of attention management / energy management strategies that are worth considering as long as you balance the downside risks.

Expand full comment

I believe there is a difference between writing concise, and writing a summary. The former attempts to cover all important details in the fastest way possible, the latter attempts to convey the general meaning of a body of text, enough so you understand the main points.

If you look at Graham’s essays, you wouldn’t call them summaries. But you would think them concise.

Expand full comment

I like this but there seems to be a a logic flaw, as it refers a video, not text.

So. The question remains: is summarizing a YT video worth it? Many times I think it is.

Expand full comment

I imagine summaries might be great when I want to decide if I really want to engage with something - but they should not substitute this engagement.

Expand full comment

The tragedy is that many AI applications lose the context in summarizing—they either focus on wrangling the facts or wrangling the fragments. So often, the story is lost.

Expand full comment

Speaking of water the human body is 70-80 percent water.

Water is also associated with Spiritual life.

The principle of Spiritual life has to do with the transformation of water, in which you free up your solid, mortal identification, and through which you become transformable, watery, alive, and your Spirit-Life begins.

The life force is water. Emotion is water. I is H2O. I = H2O, meaning the sense of independent consciousness, or the body-mind-complex, equals water.

Whatever water can discovered to be, if you do samyama on it, if you enter into its molecular and atomic domain, then there is only one H Happiness.

This earth-world watery domain is the Radiant Water of Consciousness Itself.

Expand full comment

As I see it, the use of LLMs is most beneficial when you are working within your zone of proximal development. I have received startling insights through LLMs, but those insights are only valuable if I can validate that information via at least one means that can be traced back to a primary source. Otherwise, I risk becoming a talking parrot via listening to a talking parrot. The key, as mentioned in this article, is to use the LLM to get into the neighborhood but to build up an understanding that depends on multiple lines validation outside the LLM.

If you aren't building up your own knowledge, you won't be any smarter than anyone else sitting in front of an LLM. My speculation is that those who build knowledge will be able to use LLMs as a force multiplier while those that mindlessly use LLMs will be easily replaced by the LLM itself. LLMs won't level the playing field. They will accentuate the differences.

Expand full comment

"I think one should write as much as they can with their own empiricism, their own senses, giving the reader their own characterization of life or events." Yes, thank you!

Expand full comment

"I think one should write as much as they can with their own empiricism, their own senses, giving the reader their own characterization of life or events." Yes, thank you!

Expand full comment

Before content creators making YouTube videos took over production of tutorials and documentation and information dissemination in general (because it was monetizable) the standard was text. The thing about text is it is very easily searchable, and it's also easy for someone to scan and understand the points by reading the intro the first sentence of each paragraph, etc etc. all of this is lost when content is only in videos. So I would love absolutely nothing more for AI to auto parts every YouTube video and be able to write me the cliff notes. This is the perfect way to use AI. Also I'm just going to say AI is a stupid term this is just an algorithm.

Expand full comment

This is nice. For me, I’ve a preference for YT summary, I appreciate affording leisure to the written paragraph instead, and I don’t think anything I ever write or publicly think will make it into an AIs coffers. That being said, you confirm what I’m determined to change about my own “summary-styled” writing. I know I’ve been concise —

- to put down what I think before it disappears

- bc I lack confidence in the the worth of my thought

- knowing the work involved in style & craft, I remain lazy & impatient! Don’t I want to grow?

- bc I forget my reader’s joy & investment

- hastily dismiss pleasing my reader by relying on combox engagement instead

- to remain hidden in anonymity, since details flesh out persona into *person*

But I’m bored with all that. :P

Your entry confirms a few things for me, Simon, ty. Happy New Year!

Expand full comment