Makes you wonder as well if the expressed genetic traits we can’t see are more are less different than the ones we can.
For example, does evolution have any pressure to produce those who think linguistically, vs healthy hair and skin?
810 karma · joined July 3, 2020
bram @ bramadams . dev
For example, does evolution have any pressure to produce those who think linguistically, vs healthy hair and skin?
You nailed it! Thanks for noticing the divergence!
I’ve personally found that tagging is less robust than LLM embeddings (mainly due to dimensionality), but human appended thoughts about a source — also embedded — serve even better as tags.
Example: “this is a quote about dinosaurs…” (Old way of doing things) Tags: dinosaurs, jurassic, history Query: “dinosaurs” > results = 1…
(New way of doing things) Embedded Quote: [0.182…] User Added Thought: “this dinosaur reminds me of a time i went to six flags with my cousins and…” Embedded User Added Thought: [0.284…]
Query: “dinosaurs” > results = 2 (indexes = sources, thoughts)
The "thoughts" index can do a second layer cosine similarity search and serve as a tag on its own to fetch similar concepts. Basically a tree search created by similarity from user input/feedback loops.
[1] - https://github.com/bramses/quoordinates/blob/1b9d1fadaded98b... [2] - https://github.com/bramses/quoordinates/blob/1b9d1fadaded98b...
[1] - https://github.com/bramses/quoordinates [2] - https://github.com/bramses/commonplace-bot
Whether it’s done under the auspices of Twitter Spaces or VCs promoting certain projects over others implicitly by kingmaking projects with massive seed rounds, the result is the same.
This is why I say in the post that nuance is important.
Leveraging existing LLM technologies and putting them in software where regular people can use them and have a great experience is important, necessary work. When I studied CS in college the data structure kids were the “cool kids”, but I don’t think that’s the case in ML.
The daily practice is to sketch applications, configure prompts and function calls, learn to market what you create, and try to create zero to one type tools. Here’s two examples I made, one where I took the commonplace book technique of the era of Aristotle and put it in our modern embeddings era [1] and one where I really pushed to understand the pure MD spec and integrate streaming generative models into it [2]
People seemingly desire to zealously defend the LeetCode way of solving problems for some inane reason.
Cal has applied systematic and algorithmic thinking to the messy world of productivity annd life, and teaches unique frameworks to this day about how to pursue a life that is eudamonious and well lived.
It’s also distressing that you think that mathematics research is more important than your fellow man having a fulfilling career.
I post a lot about PKM (mainly Obsidian), automation (iOS shortcuts and Alfred), slice of life posts, AI and philosophy. I also have a weekly newsletter!
Side note, if your goal is to call someone a doofus and make yourself look smarter, you should probably learn to spell it correctly.
Here ya go! (Requires Templater plug-in)
``` Week: [[<% tp.date.now('yyyy') %>-W<% tp.date.now('ww') %>]] ```
Video: Batch make videos, schedule clips for release over the week (YouTube)
Audio: I haven't used it, but I've heard good things about anchor.fm
Text: For socials, Buffer (https://buffer.com/). For newsletters, Substack and Revue both have scheduled publishing functionality.
Keep in mind, the end goal of this strategy is to work with how the internet communicates: repeated signal over time in a mountain of noise. Scheduling allows for this.
Recently, I've been more engaged on apps where you can batch thought dump and then schedule posts. This allows for:
1. community growth without overwhelm
2. (relating to 1) relationships take time, but as a soloprenuer, time is rarely available
3. batching all you want to say in (video, text, audio) format allows you to schedule "talking" and spend the rest of the week "listening"
4. (relating to 3) I reject real time communication apps unless I really trust the small community -- the stress of being on 24/7 through all concievable timezones is not healthy communication, don't fall into this trap
However, when examined with a level head and long term thinking, the set of things money can buy is much much much smaller than where it is useless.
It is in this vast domain that the only currency is knowledge, the only value in truth, the only investment in discovery.
Case 1: a few posts of mine go viral and land on the front page on HN or Reddit and I get a few to ten thousand ish views. Those days I get three to four emails from people on my contact form and we have a brief convo
Case 2: I have a few tutorials that have been linked back across the web and receive daily traffic from Google.
Outside of those two cases, most of my posts get little to no traction. My analytics are public [1] so you can see an example if you’re curious!
Unrelated Readwise questions (love the work y’all do, thank you!) - any idea when the chrome app will support multiple highlights in one go? - is there a random quote API that references all sources at once? - has the team considered adding semantic search to the app?
It has completely replaced Pocket/Instapaper for me.