Notes from 28/1/19–4/2/19
Ah this is Lunar New Year week! Two days off from the term to consolidate whatever I have and prep myself for the journey ahead again.
So there were about 100 newsletters I had to trawl through this week. Oddly I find that a lot of them tend to repeat articles I’ve seen on Facebook or from other newsletters. And out of these about 16–20 caught my eye. Some didn’t really warrant a comment, others I’m keeping them as a tutorial to try out soon.
General
This caught my eye because I never associated pottery with being a tycoon. But hey every product has it’s time.
“Once the world was out that a limited number of new vases were available to the privileged,” writes Dolan, “the price would simply reflect the idea that only people of status had real taste”
It was a brand new principle of marketing: “Deny the majority the ability to purchase art, then use their lack of means as inadequate taste.” Once an item would “grow stale” with the upper crust, he’d cut the price and market it to the wider, aspirational class…
His solution to this was multi-tiered: 1) He constantly blitzed the market with new products, and 2) He developed new ways to sell them
2. Run a Comprehensive Background Check on Yourself
Not really relevant for a Singaporean. But intersting read to see what people do when they run a background check. Even though the article is mostly on American material, it’s probably possible to automate/consolidate the datasets these companies have and create a… social credit score.
AI,China and Africa, Agriculture, Climate change, Genetic engineering, Neural interfaces, Online learning, Crisis of faith in government.
Article mostly covered the current hyper on these topics. Was hoping for something more in-depth on why these things would accelerate suddenly.
4. 9 Principles of Service Design
service design is: The activity of planning and organizing a business’s resources (people, props, and processes) in order to (1) directly improve the employee’s experience, and (2) indirectly, the customer’s experience.
Suggests mapping the users’ experience to identify pain points. Mitigate waiting (oh this happens too much). Follow-up with your customer to make sure things are going fine. Pre-empt customer needs.
Computer Science
Connected group of Qubits can provide much more processing power than binary bits.
Due to superposition, qubits can represent numerous possible combinations of 1 and 0 at the same time. This allows them to compute a vast number of potential outcomes simultaneously.
Entangled qubits result in the other qubit instantaneously changing state in a predictable way even if separated by long distances.
The quantum state of a qubit is extremely fragile resulting in them to become ‘decoherent’ which is to decay.
2. AlphaStar: Mastering the Real-Time Strategy Game StarCraft II
Spent an entire supper discussing with friends about how DeepMind seems to be overselling their win. Contentious points were: mean APM for AlphaStar is actually much higher as it was able to sustain high APM when needed, sheer amount of power required to run the AI, cost of training, AI cheating in the game, that the professional players weren’t playing their preferred race.
But all that considers, that was pretty amazing. My theory is that AlphaStar is just a really optimised 3500MMR player because that’s what most of the dataset consists. Though I should really check this theory.
Potentially we might see some sort of AlphaStar zoo optimised to defeat each pro player.
I’m very seriously considering doing my decision theory case study on this now!
3. A self-learning, modern computer science curriculum
Keeping this here because it’s so well curated
Blockchain
Argues that smart contracts form some sort of foundational layer and thus must be extremely stable for more applications to be built on top of them. That dependencies must rely on stable code/ components in the architecture rely on foundational components that are stable.
Not sure what this article is trying to propose here. I was expecting something about loop invariants.
2. Incentivizing Competition in Blockchain Foundations
Have a competition! Get them to compete! Don’t just rely on grants!
3. Comedy crypto: How Holland’s most popular comedian is beating ticket scalpers
Puts ticket owners and state changes of the tickets on the Ethereum blockchain. Registers proof of the ticket on the blockchain via customer’s mobile phone. GUTS (the company) thus has control over the life cycle of the ticket. Runs on the GET cryptocurrency.
Time to take a closer look.
Somehow still not convinced you need your own token — but hey this could be something to do!
4. Spacechain
Somehow never heard of this.. but wow okay Qtum-based blockchain nodes are in SPACE!
Blockchain-Ethereum
Supporting the Goerli testnet with PoA nodes, acting as a faucet to onboard validators onto the Eth2.0 Prysm testnet.
Mainly waiting for sufficient validator deposits to start the beacon chain.
This is probably quite exciting because it’s been a long-awaited project. Even more excited to see what sort of interesting behaviours validators might pull…
2. Why Ethereal Summit is The Most Unique Event in Blockchain
It sure looks like the most wholesome thing that’s ever happened to blockchain. Really like the diversity of the summit.
Life Optimisation
Process goals are not about what you get done, but about what you do. You reach your process goals by behaving in a certain way, not by reaching some bar of quantified outputs — your behavior is the performance. It’s about how you spend your time instead of what exactly you produce during that time…
If it is a taking-care-of goal, you’re going to need process goals. Your aim should be to produce high quality work during the hours you put in.
If it is a making-sure-that goal, you can pull out those traditional to do lists. Your focus should be on output maximalization.
Knowing why you do what when. Keeping the objective of your goal in mind when you are doing it. Not everything can be measured with numbers only.
Not sure what the benefit is here…
3. The Definitive Guide to Contributing to Open Source
Will definitely adapt this for OpenSUTD
