Building something new. Self-taught designer and dev. Product guy, formerly of @revolut.
Joined January 2009
- Tweets27.7K
- Following1.9K
- Followers2.6K
- Likes21.4K
Andric retweeted
Much decision-making is tacit or qualitative. Making those parts more legible is the first step towards observability & steerability. If given the right primitives and data, AI systems could help us codify this nebulous process, maybe using our inputs/outputs to derive our decision-making function as a starting point.
But how will it be trained to do this, and on what data? As of yet, there are no good primitives or tooling for codifying our own OODA/feedback loops: to express + record context (inputs to decision-making function), principles/constraints (the function), and results (function output), plus the meta work of process evaluation & refinement.
I don’t really want to prompt an AI system.
Instead I want it to passively watch what I do as I do things, talk to me to understand what I’m thinking as I do them, and progressively learn from me.
Over time, it should understand tacitly what it’s like to make certain decisions, and then do them.
Like my personal apprentice.
Just like training an apprentice, a true AI agent should amplify your abilities to the limits of how much compute you can buy, so that you can do a lot more with the same amount of time (which is true productivity).
Current “AI agents” aren’t like this at all. They do things on your behalf in ways that their developers decided they should do.
An AI agent should instead do things exactly how I would do them.
Do as I do, not as I say.
Andric retweeted
re: chiang—
it's futile to argue that a paintbrush, because it does not experience humanity, cannot create art. It makes two category errors: one, mistaking that art is created by tools rather than artists; and two, that art is christened by construction, rather than culture.
Andric retweeted
I've developed a lot of plugin systems, and the OpenAI ChatGPT plugin interface might be the damn craziest and most impressive approach I've ever seen in computing in my entire life.
Andric retweeted
Good observation here. More info coming at us = more need for tools to parse it = more 'interpretive labor'
I'd add that info we're consuming is also shorter and more staccato than books of yore, makes getting into deep context (10pgs in insights) harder which = more labor still
Andric retweeted
Tech is only interesting long term, like over decades, if you're loyal to the tech itself. As in, the slowly uncovered laws and principles of how technology works. If you get trapped by the people-level loyalty circus, ultimately you'll lose whatever tech instincts you have.
Andric retweeted
This is actually a good test. If a little friction stops you you’ll end up with twitter “alternatives” that offer distinctions without differences. It’s not a coincidence that the 2 serious options, mastodon and farcaster, have friction.
This quoted post is unavailable.
Andric retweeted
Ethereum gas fees aren’t a weird artifact of blockchain technology. They’re just legible real transaction costs priced in energy dollars. Any tx that looks much cheaper hides things like petrodollar externality costs, crime, employment uncertainty, contract break penalties etc
Andric retweeted
We've allowed the banner of "innovation" to be hookwinked by people who lack the ability to find creative solutions to the problems before them.
We've allowed "first principles" to be distorted into an utter lack of principles.
And all it is is intellectual laziness.
Andric retweeted
Take ideas that come from larger grand-unified theories but are often deployed in stand-alone isolation, like “raise the minimum wage” or “lower interest rates” … that’s theory-shaped debris.
Bits and pieces of collapsed theories nobody really defends in complete form anymore.
Andric retweeted
This is why I can’t really bring myself to care about NFT culture wars. That feels like people in Europe arguing about whether Spain or Portugal or Britain will gain power from New World. In the New World itself, there’s immediate concerns like moving pianos.
Andric retweeted
The thing about NFTs is that they are the opposite of finished products.
The crypto kids know they are immaterial.
They know there is no competitive market space left for the things they want to value.
4/
Andric retweeted
"Imagine if every application on the Internet by default had a zero-downtime, publicly exposed API"
people sometimes ask, "why is building fully on-chain interesting?"
one of the best articulations of this is tucked in the footnotes @saffronhuang's recent piece on crypto-native governance schemes (see it at 0xparc.org/blog or at baby.mirror.xyz) (1/n)
Andric retweeted
TLDR: Web3 has one real function: to kill web2, so that web1 can live.
A major problem in the web3 debate is the assumption that web1 was a phase that began in 1993 and ended in 1999. This confuses a whole philosophy and culture with a fixed timespan.
1/12
Andric retweeted
New movements develop new aesthetics to distinguish themselves from their predecessors.
As these movements gain influence and cultural relevancy, others coopt their aesthetics to gain the benefits of association with the movement without needing to adopt the values or practices.
Andric retweeted
Every now and again, someone criticizes a project by saying something like,
"why are you having eggs for breakfast when you can have pizza for dinner?"
and all it reveals is how trapped the speaker is in a scarcity mindset.
one does not invalidate the other.
you can do both.
