- The Limitless Playbook
- Posts
- #008 | Be a Specialised Generalist
#008 | Be a Specialised Generalist
Hi, I’m Ryan! You are receiving this because you have signed up to my weekly newsletter for Natural Language Processing (#NLP365), Entrepreneurship, and Life Design content!
Hey friends,
This week’s is all about the strength of being a specialised generalist. Specialised generalists tend to be great problem solvers and have an extremely high executional power. Your executional power is simply your ability to get things done. And by “things”, I mean, in general and not bound to any specific domains. Why should we strive to become a specialised generalist? and why does the world only rewards specialists? Let’s see what the book Range can teach us :) Here’s a snippet of my Level 4 notes on the book.
The world is a ‘wicked’ environment
There are two types of environment: Kind and Wicked. Kind environments are environments where patterns are repeated over and over again and feedback is usually rapid and accurate (e.g. chess). In this environment, deliberate practice and the 10,000 hours rule can be apply to achieve early specialisation. In a kind environment, you improve simply by engaging in the activity and aiming to get better.
In Wicked environments, the rules are often ambiguous or incomplete. There may or may not have repetitive patterns and feedback are delayed, not obvious, or inaccurate (or all of them). When you approach a wicked domain with narrow specialisation (that works well in the kind domain), it can go horribly wrong as you tend to rely on familiar patterns.
Human's greatest strength is our ability to integrate broadly, the opposite of narrow specialisation. Successful adapters are great at applying knowledge from one area to another.
Innovation / Problem Solvers
Outside-in thinking is when we find solutions from people that's outside the domain problem itself. The further away the problem is from the solver's expertise, the more likely that they are going to solve it. This gives us a higher chance for serendipity moment and it counters the Einstellung effect, where experts have the tendency to employ only familiar techniques eve if there's a better method out there.
Knowledge is a double-edged sword. With it you can do and build things but it also blinds you to other things. There's also something called "Undiscovered public knowledge", where there exists public information that are still unconnected and waiting to be discover. The more specialists create information, the more opportunity exists for outsiders to connect public knowledge. Essentially you can push forward by looking back. You can apply old knowledge in new domain. After all, a lot of useful knowledge are still not connected.
This week I finished reading:
Range (20th Feb - 23rd Feb 2021)
Never Split The Difference (25th Feb - 27th Feb 2021)
Total: 16 / 26 books | 2 / 26 level 4 notes | 1 / 12 actions
🔥 ANNOUNCEMENT 🔥
Join Zeroton’s ACTION-FOCUSED Slack Channel — My friends and I are currently working on a “knowledge”-sharing platform that converts books into actions, with the mission to fight the problem of information overload! Check out Recommendation(s) of the Week section for more details.
❓Question of the Week
Are you a generalist or a specialist?
People often confuse lack of specialty with being a generalist. That’s simply not the case. To determine if you are a generalist, imagine you walk into a mansion with hundreds of rooms, each represents a different domain. How many rooms can you walk in, hold conversations, and contribute to the room’s development?
A generalist can be someone that has a high level of expertise in a specific area (but less than a true specialist still) and also has breadth of knowledge that covers the different areas of development in life.
Share with me your thoughts by replying to this email 👻 👻 👻
🐦 Tweet of the Week
I admit to judging books by their cover
— Elon Musk (@elonmusk)
11:10 AM • Feb 25, 2021
💡 Quote of the Week
You’ll often find that people’s declared preferences—what they say they want—are far different from their revealed preferences—what they actually do — Hooked
🔥 Recommendation(s) of the Week
Join Zeroton’s ACTION-FOCUSED Slack Channel — My friends and I are currently working on a “knowledge”-sharing platform that converts books into actions, with the mission to fight the problem of information overload!
We are in the process of looking for STRONG ACTION TAKERS, who believes in our mission and also wants to self-improve themselves through non-fiction books, to join our slack community. We want to connect and learn more about your personal development journey. We are aiming to keep this group small (20 - 25 people) to ensure that the quality of interaction is high.
Through the slack community, you will have access to the following:
Early access to Zeroton’s 100+ of Actions from 30 best-seller entrepreneurship and personal development books
Weekly group zoom calls with myself and other like-minded people to discuss actionable insights from different non-fiction books. We will keep each other accountable and share personal experience in testing different actions! Essentially, we will grow, learn, and succeed together
Early access to the Zeroton platform
To join, please reply to this email with “I am interested in joining Zeroton’s slack channel” and I will reach out to you with more details. Looking forward to connecting with you guys on the slack channel!
🔦 AI Research Paper(s) Spotlight of the Week
Proposed a novel framework for inferring missing entity type instances in KGs using entity type instances and global triple information and two mechanisms:
Mechanism 1. Missing entity types of an entity can be found from other entities that are close to the entity in the embedding space, using local typing knowledge
Mechanism 2. Missing entity types of an (head or tail) entity can be inferred from the types of other (tail or head) entities through their relationships, using global triple knowledge
Neural models have been driving SOTA results in NER but it requires large amount of annotated training data, which it's not readily available for all languages. This paper proposed an unsupervised cross-lingual neural NER model that can do transfer learning from one knowledge to another without relying on any labels in target language, bilingual dictionary or parallel data.
🎥 This Week on YouTube
That’s it for this week! I hope you find something useful from this newsletter. More to come next Sunday! Have a good week ahead! 🎮