> Programmers continuing to become more expensive, computers continuing to become less so. Therefore, the smart bet was on making those programmers more productive EVEN AT THE EXPENSE OF THE COMPUTER!
I used Blot for about 5 years for Second Breakfast. Its ease of use got me started blogging. Very cool app/service, highly recommend.
I had it strung up with RSS to Mailchimp to auto-send new posts to a mailing list. Recently just switched to Ghost to make that more integrated, we'll see how it goes!
I believe Eliezer uses this as a definition of intelligence, but it also works as a definition of agency: to paraphrase, a system that acts to attain or impose a pattern, criterion or constraint on future states of the world.
A crucial metric, IMO, is the degree to which paths of action to these criteria extend from the agent. For instance, a Spot robot acts to maintain the criterion of upright position on its future stance by action of leg movement. This only affects the robot directly through short plans and so is relatively harmless. In comparison, an ASI running on an AWS datacenter may impose the criterion that the datacenter continue to exist, through long chains of action involving the eventual death of all humans intending or posing a chance to destroy it. That would obviously be quite a lot worse, but I think the example illustrates how "imposing a criterion onto the future" captures the essence of agentic behavior at various levels of power and danger, without pulling in any unnecessary detritus such as "consciousness", "will" or "emotion".
I don't have a good definition of agency, but I do think agency is required before consciousness. I believe consciousness is the recognition of how your agency influences yourself and the external world. Having some mental model of the consequences of your own agency.
The dictionary definition is just the ability to take action over something. It sounds like you are using a definition that relies on consciousness which is a much more complicated and vague concept.
For example, I think most people would agree that my pet cat has agency. It can go wherever it wants in my home, eat whenever it wants, sleep whenever it wants, etc. Whether it has consciousness is a much more controversial topic. Basically everything living has agency. Even my houseplant will direct its leaves toward the sun, but few would argue it has consciousness.
Haven't thought of this. Couldn't you just give a model a way to constantly "chew" on something? Maybe an ever-ending loop of some sort of prompt stimulation?
An even more fun experiment will be to have two models running in perpetuity, constantly talking to each other, but constructed to act as though they were two sides of the same model.
Do you have an API where you can get audio clips back reasonably quickly? Like if I wanted to use this in a voice support bot, could I send a text blurb to an API and fairly quickly get back an audio file?
Teaching the history and story of the scientists who made different genetic discoveries, and along the timeline they made those discoveries, made everything make sense.
I disagree that there are too many people. Have you read much about the Malthus/George debates? Malthus has now been wrong for 200 years (and counting).
We can have population growth and still solve the impending environmental disaster.
On top of that, a declining human population results is an increasing young:old ratio. Fewer younger working age people supporting an increasingly older population. Socially and economically that will be a disaster, which will make solving ecological issues significantly harder.
We can have growth, but the growth rate must decline over time. For example the current growth rate of world population is around 1.1% per year. It is easy to see that it cannot remain above 1% for more than 12101 years (assuming FTL travel is not discovered).
Proof: Right now every living human is either on Earth or close to Earth. Without FTL it follows that 1 year from now every living human must be within 1 light year of Earth, 2 years from now every living human must be within 2 light years of Earth, and so on. 12101 years from now every living human must be within 12101 years of Earth.
If you calculate the volume of a sphere of radius 12101 light years centered on Earth and divide that volume by what the population would be after 12101 years of 1% annual growth you get 0.04 m^3 per person.
But the volume of an average human is ~0.06 m^3, which is greater than 0.04 m^3.
At 0.1% annual growth it would take a little over 127000 years to run out of room.
At 0.01% annual growth we've got 1.34 million years.
Of course the actual limits are much lower because the above is assuming that we can pack humans so that there is no space that is not occupied by humans which we cannot do because, among other things, (1) our shapes don't fit together perfectly to allow such tight packing, (2) there isn't enough mass in the 12101 light year sphere to make that many humans, and (3) we would need to use much of that space for the infrastructure needed to support humans.
Humor aside, it's a depressing worldview to me that each marginal human is a net negative. I take the opposite view, that each human is a net positive - and we need all the brainpower we can get to solve all of our problems.
We could solve a lot of our problems right now if people were prepared to change how they live and accept a decent amount more responsibility (e.g. sort waste by hand, etc.). Scaling up problems typically hasn't helped, and you get novel issues at large scale (e.g. parasites and diseases thrive due to monoculture in huge farms). As Bill Burr noted, we could all drive tanks to work if there weren't as many of us. It's cynical, but there's more room for largesse at small scale.
Pyramids were built with slave labor. Poor sanitation lead to the black plague. Imagine if they all stopped having kids, because they thought Egyptian slavery would never end and the black plague was insurmountable?
Bottle | Founding Engineer + Head of Engineering | NYC preferred | Full-time
Bottle helps local food brands sell to all of their customers - pickup orders, delivery, gifts, on-demand, pre-order, wholesale, subscription - in a single unified checkout flow.
We bootstrapped for 5 years, raised money a year ago, and are in growth mode now. We're a tight product team - just four full-time engineers and no dedicated product managers yet.
We're hiring founding engineers who want to be on a tight-knit team, take on really hard challenges in a huge market, and consistently ship beautiful software.
Offering significant equity (1-5%) for high performers. Email me directly, will at bottle.com
Simple but brilliant insight.