You can have an anonymized dump like you'd normally do and then branch it. This allows spinning up environments in seconds and without the disk footprint of a new replica or dedicated DB.
The privacy / security constraints stay the same whether you are branching or not.
Anonymizing data is the biggest part of such a workflow. Most prevalent use case that requires production data is for debugging. I guess there is some value in branching non-prod databases for feature development.
Most security teams do not allow prod data in non prod environments, anonymized or not l.
Product is not the same as code. We code to build a product, sure, but I think the author means they are interested in designing the product to solve users problems (a.k.a UX)
This is the future of applications. Still not sure if model providers are the ones to do it. I think of LLM as infrastructure and I can build apps on it in a "general" way. Not the bespoke wrapper apps that are proliferating today, but LLM as a native interface to build(and use the app).
I agree. I don't know either company, schema builder is a very common feature in many data platforms. Nested or otherwise. Neither is claiming this is a big deal though.
Congratulations! Do you compete with unstructured.io?
And how do you think about ever-improving models from Google etc? (The doc extract API, AWS has something as well)
Nothing is essential to the universe, the universe does not need us. We need it exactly as it is today, and that is it. Everything else is stuff we made up to understand the universe weakly, or to cope with life. This is not nihilism, but meaning to life is meaningful only in the context of humans and has nothing to do with universe needing us.
This is only true at the physical level. At another level, the only thing that “exists” is a mind. If all you have is a bunch of rocks floating through space with nothing to perceive it, the universe is indistinguishable from emptiness. Experience lives at the intersection of instantiated reality and thought / perception. You need both.
You can also imagine travelling along the axis of an idea or an archetype through time and space. For example, the idea of lovers or warriors or something. Each instantiation of that idea in someone exists along that axis. The idea can only come into being inside a physical reality or simulation. But the idea itself is eternal. The idea of the number 1 doesn’t “need” the universe.
It's a nice thought. If an idea could be encoded... that each persons idea of a concept, viewed in the right dimensionality, has a rough, similar outline. At least, that's my interpretation of your idea. :-)
Your comment either triggered, or made me notice an ongoing shift in my worldview. Thank you!
> But the idea itself is eternal. The idea of the number 1 doesn’t “need” the universe.
Ideas are mental concepts, thoughts, or notions that exist in the mind.
As far as we know, there’s no evidence of mind-independent ideas. We can only access ideas through our minds, so we have no way to verify they exist without minds.
Is an LLM "a mind"? They certainly seem to encode brain-independent ideas.
> We can only access ideas through our minds, so we have no way to verify they exist without minds.
Its symmetric. We can also only access the physical world through our minds too. We have no way to verify it exists without a mind. Actually we have no way to verify it exists at all in the form we perceive it - since it might just be a big simulation.
In that sense, I have more direct evidence of my ideas than I do the world, since I can perceive ideas directly. (I say "my ideas" not "our ideas" because I can't tell as easily if you exist.)
If you see a large enough, dangerous-looking enough animal in person and up close, you will respond physically without a single thought or idea. If you didn't, evolutionarily you wouldn’t exist.
This is because you are parallel systems. Ideas are not primary experience.
> We can also only access the physical world through our minds too. We have no way to verify it exists without a mind.
The physical responses that could save your life are pre-mind, and testably extant within you.
> I have more direct evidence of my ideas than I do the world, since I can perceive ideas directly.
Also, and perhaps to be considered separately, you have no evidence that your ideas are directly perceived, only that they are qualitatively different from what you have been educated to second-order understand as sensory input.
> The physical responses that could save your life are pre-mind, and testably extant within you.
Yes you’re right. There’s a mountain of things that we do that we don’t have conscious control over. There’s multiple ways to think about that.
First is that we simply aren’t masters of every corner of our minds and bodies. But so what? Nothing I said requires us to have full control or full knowledge of ourselves. The existence of reflexes doesn’t directly counter any of my claims.
Another way to think about it is by changing the definition of self / mind to only include the things in our direct perception and control. After all, the boundary of self and world is incredibly fuzzy at the best of times. Are your fingernails “you”? If I accidentally touch something hot with my hand, “I” don’t choose to flinch my hand back. I could consider that reaction part of the external world, not part of me. Interestingly that definition implies self mastery makes “you” bigger. When children stop having tantrums, they bring their emotions under control of their mind. In the process, the part that is them gets bigger to encompass their newfound control.
My physical reaction to tigers also in no way disproves the simulation hypothesis. None of us can directly perceive a tiger.
> you have no evidence that your ideas are directly perceived
Interesting! What do you think we do have evidence of directly perceiving? Surely not a tiger.
I perceive thoughts. I don’t perceive my sensory data - but I do perceive my interpretation of my sensory data. (I don’t see pixels or sound samples. I see blue chair, door, ... I hear dog barking (low pitch), I feel headache feeling, etc). I think that’s good evidence that a pattern of “chair” and “dog” exists in my mind. And I can think about the number 4. When I do that, something happens in my mind. I don’t think I’m perceiving “four” with my sensory system. But the thought has to exist somewhere, right?
> My physical reaction to tigers also in no way disproves the simulation hypothesis.
This is a big conclusion to state with no justification and no definitions.
You seem to equate simulation hypothesis with solipsism. Are you actually equating these?
> I perceive thoughts. I don’t perceive my sensory data — but I do perceive my interpretation of my sensory data. (I don’t see pixels or sound samples…
You are missing the fact that in your view all of that sensory information, and all of what you learned about how it works, is simulation. You cannot rely on the fact that pixels are something you have observed to tell you that pixels should be your fundamental light sensory units.
It seems also to be mistaken reasoning that a simulation should emerge from present-era computing and electronic sensing principles.
> And I can think about the number 4. When I do that, something happens in my mind. I don’t think I am perceiving “four” with my sensory system. But the thought has to exist somewhere, right?
Yes, but think about how you got there. You thought about 4. How is that critically experientially different from looking at the door handle. Or listening to the rain.
To answer your question, without having considered the answer much: I think the only things I actually do are perhaps observe, explore and exercise volition.
Their instantiations are based on either some physical hardwiring (instincts) or cultivated though observations. If an idea has the appearance of being eternal, it's because either there's a process in place for that physical structure to be replicated though time, or because new minds are making essentially the same observations.
I suspect even some of the most "eternal" of ideas are less so than you might think. If we were capable of communicating with a human from 100,000 years ago, in detail, how many identical ideas would we share? I think it would be very few, given that many people's ideas already differ subtlety even when they live in similar environments.
If you crushed all the atoms in the universe into dust, would the number 2 also be destroyed in the process? I don't know if the mathematical idea of 2 exists in the physical world at all. I don’t know if that’s even a well formed question.
Obviously humans perceive 2-ness. And if you crushed all our minds to dust, nobody would remember 2. But would that really destroy the number? If an alien civilisation in another universe counted things, I think it might be the very same number 2 that they use to count.
Numbers show up in a lot of different cultures. If we all independently reinvented / rediscovered the same thing, that’s interesting - don’t you think?
Does convergent evolution prove that a structure like an eyeball exists in some abstract realm, or does it mean there's a process which - given a similar environment - will produce that structure over time?
It's not hard to see why the same useful inventions emerged in similar environments.
The only way we see ideas manifest is in a physical world, and via minds (predominantly human at that). While ideas are not made out of atoms, they exist within atoms (brains, computers, sound, books and so on). We can imagine ideas exist on their own independent of their manifestation - but its just that - an imagination.
But plenty of ideas can’t be made manifest in the physical world. Like the idea of infinity doesn’t - and can’t - exist in the physical. Yet it’s still a useful idea in mathematics.
And again, the physical cannot be perceived except through a mind of some kind. To play devils advocate, we can imagine the universe existing without a mind to observe it - but it’s just that - an imagination.
I agree Amazon should pay taxes. But this bill is not the way to make such companies pay taxes. It will kill competition and startups along the way. That is the crux of the issue.
Edit: Looks like Amazon did indeed pay 15B+ federal taxes in 2024 (excluding sales tax etc)
Incidentally, should we really count "income taxes" as something "Amazon pays"? Amazon doesn't pay income taxes. Amazon employees do. The fact that Amazon conducts the transaction via withholding seems irrelevant. It's the employee losing the money.
i generally think they should be counted to the company as a sales tax. amazon is losing the money, because theyre paying it to the government and not the employee, ao they need to increase the pay accordingly if they want the employee to have a certain amount.
reply