As a customer of Konfig, I am super bummed about this. I think we may even have been their first customer (referenced in the init commit). Dylan and Anh-Tuan solved a number of real problems for us, and they did it spectacularly. They essentially provided a one-stop-shop for generated SDKs and beautiful hosted docs/tutorials that would call our API and immediately allow users to start playing with real data. Prior to Konfig, we were hand-writing a hodgepodge of SDKs to wrap our API and using ReadMe to host the docs (which never quite worked interactively with our API due to request signing that we require on our API).
Tons of respect for these guys and I wish they'd found a way to make it work.
Also, for anyone finding themselves in a similar situation as Konfig (a product that (some? most? all?) customers love, but revenue not scaling as needed), please consider charging more. We probably would have paid 2-3x what Konfig was charging us and more as we grow, but they never asked and never built in any usage-based cost scaling (like $X/month/SDK, $Y/month/demo, etc).
I got to know about konfig from your product itself. The docs were(maybe still are?) so good that I had to check what was behind it. It really is better than swagger/redoc.
There are multiple open-source tools backed by large companies to do this. I use one of them for my sdks. So I can really understand how they could struggle to sell it, it's a hard sale to sell someone a commercial product that does pretty much the same as free open-source projects. I was tempted to check it out but as bad as it sounds even with it being open source and free I couldn't find a reason to spend the time over just using what I have now.
The product in TFA is a wrapper around the tool I linked, plus some stuff for docs and so on.
Wrapping free software is mainly viable towards enterprise customers, where the thing you're actually selling is SLA and a certain level of insurance and sexiness on middle manager CV:s and so on.
I was looking forward to that as well, but Siri refuses to read ChatGPT's responses to me...which makes it substantially less useful. It's just as frustrating as being told "sorry, you'll have to unlock your phone if you want me to play something on youtube music". Let's hope they iterate, I guess?
For speed of light, a more direct analogue (though your example still makes sense) would be what is often called the “speed of sound” in numerical methods (like the finite difference method, aka FDM), which is often closely related to both temporal and spatial discretization schemes and can lead to numerical instability (cf. CFL condition, Von Neumann stability analysis, etc). It’s also related to things like stiff problems. It essentially has to do with how fast computational information propagates through the simulated domain. Note that it is still called “the speed of sound” even when you are not simulating the wave equation! (At least by my profs)
The simulation framing tends to draw deterministic connotations. I think when people think of a simulation or perhaps physical interpretations, they imagine life as happening to them, rather than through them. I think it’s easy to perceive it one way or the other, and what you seek you will find.
But something like ‘lazy evaluation’ could provide a bridge between the two views. A choice is made AND a physically compatible path is observed as if it were always so from the past to the present.
Personally, I think of the universe is alive with choice and we do get to participate.
PS. I kept a copy of Kip Thorne’s Black Holes and Time Warps on my nightstand as a kid - it filled me with fascination. Thanks Kip!
> The simulation framing tends to draw deterministic connotations.
I bristle at this because there is obviously a massive amount of probabilistic programming methodology and stochastic simulation techniques out there, but at the end of the day in terms of connotation you are probably correct…
I think there's a case to be made that, if the simulation hypothesis is correct, we are god's dice[1]. We exist as agents to introduce randomness in the form of free will or enough chaos to ensure non-determinism, at least from the perspective of whatever force built this place.
1. The laws of the universe require unimaginable power, knowledge, and precision to implement. Its Creator must have these traits.
2. It doesn’t fail. Everything humans build requires maintenance but still fails. That this universe runs with a 100% reliability rate shows its maintainer’s power, knowledge, and perfection. It also shows He is sustaining us every second. He’s thinking about and caring for His creation.
3. Humans appear hardwired to seek who God is. So, the Creator wants us to look for them.
4. Even as children, humans have basic notions of love, justice, and are relational. The Creator is a relational, moral being who wants them to think like that.
5. All life passes through genes that carry some appearance and behavior. The higher creations raise their offspring. The Creator designed us to bear children who we teach. We’re creators in a way, too.
6. Many claim the Creator reveals Himself to them. One had objective proof, power, and outcomes that back that up (eg GetHisWord.com). The Creator has purpose for us which He helpfully shares. Jesus also lived it, too, in the same flesh before dying for our sins.
Just these six suggest that, in contrast to gaming, our world is run by God who has power, omniscience, morals, and a relationship with His people. So, that says the way to win is to be close to Him (Jesus) while playing the game by His rules. Also, to enjoy learning how it works which is endlessly deep and fun. :)
In contrast, a game engine or computer doesn’t compare mostly because of how the universe works perfectly despite billions upon billions upon billions of interactions. It can’t be overstated how unlikely that is based on all observations of reality. It’s truly astounding.
If you are willing to argue in good faith (no pun intended), I'd recommend for you to read Spinoza. Spinoza builds on your argument number one and argues that there can only be one substance, and this substance is God. In a nutshell: God is everything that exists, we do not exist outside of God (we are "modes" of God, if I remember correctly). Spinoza also argues that by virtue of being the only substance, God exists necessarily and does not have a choice.
The implications of this logic create problems for the Judeo-Christian stance. Absolute morality goes out of the window and a few other things with it as well.
It might be interesting to read. It's not far from my original guess about these things.
The logic wouldn't create a problem for us due to the weight of our source, the Word of God. Whatever counters it would need perfect character, prophecies that came true, miraculous power, historical evidence, and global impact on most people groups. Then, his followers would have to experience similar things on top of transformed lives. If not, his views remain pure speculation with nothing backing them like most religions and philosophies. Not threatening at all. :)
That is why I prefixed my previous post with "in good faith". If you postulate that your speculations carry more weight without solid logical reasoning, that is not good faith to me.
Granted, that style of reasoning also has a long tradition in philosophers like Descartes, Berkeley, etc. Descartes famously postulates that "God is not a deceiver", and that we are dealing with a benevolent God. You make the same assumption. Back then, there had to be a God, because the church would have showed people how the afterlife looks like pretty quickly. I don't understand what necessitates such a stance today.
In any case: as long as you argue from the conclusion backwards, we can spare some ink and leave this be.
“ If you postulate that your speculations carry more weight without solid logical reasoning, that is not good faith to me.”
I thought my original comment had a link. We have a huge weight of evidence of various types to support God’s Word being from God. There’s usually more types for that than most beliefs people express on HN that are accepted. So, I start with that as a foundation much like proof assistants build on a core logic.
To test your assertion, we can do simple comparisons that Christians often do to justify their beliefs. For instance, you equated our use of the Bible to Descartes stating an opinion. Did Descartes live a perfect life, claim to speak for God, and perform miracles to prove that? Did he come back from the dead? Do his followers experience unlikely transformations and life events in response to praying to Descartes? Do they get healings in the hospital verified by doctors by asking Descartes to heal the person? Would the people I’ve seen who were miracle healed have done better with Descartes?
When a philosopher or scientist counters Christ or His Word, we can just go down the list to find they don’t come close to refuting them. Christ wins the trustworthiness competition. Then, we trust Him based on that.
I would consider switching sides if the others met the same criteria. They’d have to claim to receive visions from God, their predictions come true precisely, work miracles, come back from the dead, have perfect character (trustworthy), and I’d have to get promised results following them. If not, “let God be true and every man a liar” when they contradict.
I find real science doesn’t contradict my faith, though, since it’s a pursuit of truth which God wants us to pursue. Most of it is OK or it doesn’t matter if it’s right or wrong. I can enjoy it all. :)
> 1. The laws of the universe require unimaginable power, knowledge, and precision to implement. Its Creator must have these traits.
Since you are adding these observation in the context of the universe being a simulation none of the above need to be true. The one who designed the universe did not necessarily have the knowledge to implement (code the universe), build the hardware that it runs on or built the power source that is powering it all.
Why did you jump to the conclusion that it is only one person/being doing everything?
> 2. It doesn’t fail. Everything humans build requires maintenance but still fails. That this universe runs with a 100% reliability rate shows its maintainer’s power, knowledge, and perfection. It also shows He is sustaining us every second. He’s thinking about and caring for His creation.
How do you know that it doesn't fail? What would failure even look like? Wouldn't something like the heat death of the universe signify its failure?
How do you know that the universe runs with 100% reliability?
Why mention that everything that humans build requires maintenance when even the universe, by your own words, requires maintenance by its creator?
> 3. Humans appear hardwired to seek who God is. So, the Creator wants us to look for them.
Humans appear hardwired to seek explanations to phenomena. So much so that when they can not logically explain a phenomena they will make up an explanation. Isn't this more logical than your statement, if not, why?
In your point 4 and 5 you are just saying because some facts of the universe thus god.
> 6. Many claim the Creator reveals Himself to them. One had objective proof, power, and outcomes that back that up (eg GetHisWord.com). The Creator has purpose for us which He helpfully shares. Jesus also lived it, too, in the same flesh before dying for our sins.
Good observations. I’ll try to address a few of them.
Re reliability. We know what we can observe. We see the machinery doing billions of billions of things with perfect consistency. The amount of interactions required is staggering. Whereas, humans quit trying to prove code correct after many 10,000 lines.
The being is either keeping it running without failures being possible (God’s power) or is doing the equivalent of preventative maintenance. Both are logically possible.
Re hardwired. That is a good hypothesis. I’d counter along C.S. Lewis’s lines that some things we seek aren’t merely imaginative. We have a strong, unique urge for specific things we need or are critical to our species, like food or sex. Those urges are tied to objective, real things. Our urges for God, love, and justice are just as strong and global. That’s because we’re designed for it.
Re: how do I know. Well, there’s multiple forms of knowledge. They include evidentiary (eg historical), phenomenological (eg sensation of pain), logical, empirical, and revelatory. That last one is important because we can’t really know anything outside the universe unless told by those outside it. That God chose to reveal Himself to His creation solves some big problems we could never solve without that.
“ Did not share it with me. Why?”
He did. He works through people who carry His message of reconciliation, the Gospel. It was in the quote:
His Word says God draws us to Him, faith is a gift, it comes by hearing His Word, and (per Jesus) His sheep hear His voice. If people are humble, and truly seeking Him, He supernaturally tells them the message is true. Convicted of their sin, they have a choice to make that determines if they face eternal punishment or receive the mercy of eternal life.
God could’ve made it hard. He could’ve required people to be geniuses, know important people, be star athletes, or live perfectly on their own. Instead, Christ paid the price for our sins. Then, gave anyone who acts on a simple message to be saved. Anyone who shares that message might pull others out the fire, too.
Believe, repent, and live for and like Him. He’s worth it. Hell’s not.
You didn't say that about either the claims in the OP article or prior submissions with less proof of what they say. Your reaction might be emotional, not logical. I encourage you to consider it logically.
An easy test would be to find human machines that have 100% reliability without design or maintenance. Then, strong arguments that they'll remain that way for thousands to billions of years with no intelligent intervention. If you wouldn't believe that, why would you believe an atheist that says the universe's superior machinery was likewise undesigned, is unmaintained, and is accidentally perfect? That goes against everything we've observed.
We also have a guess on the computational requirements which are mind-boggling. Just simulating tiny molecular or quantum systems with total accuracy, if it's even happened, takes a ton of computation. Every sub-particle or force in the entire universe running at the same time is unimaginable computation. Plus, the Creator makes them all work exactly as desired with the same, observed patterns. We can't get cellular automata to work like we want but the Creator is directing a universe of simple interactions to cause specific behavior, global and local?
God is the best theory for these observations. That also makes Him the most interesting of all beings in existence. Since we know Him, we also worship Him on top of it. That's acquired through non-scientific means, though. He supernaturally imparts it to those who believe in the Gospel of Jesus Christ. It's more phenomenological but indirectly observable.
> An easy test would be to find human machines that have 100% reliability without design or maintenance. Then, strong arguments that they'll remain that way for thousands to billions of years with no intelligent intervention. If you wouldn't believe that
This very much might be a failing, short lived universe. Why do you think that it isn't? Keep in mind that when we run simulation we do not usually run them in real time where one of our days equals one day in a simulation.
> If you wouldn't believe that, why would you believe an atheist that says the universe's superior machinery was likewise undesigned, is unmaintained, and is accidentally perfect? That goes against everything we've observed.
It doesn't go against anything we observed. We did not observe that the universe is a machine nor that it is perfect. All that we observed is that the universe is as it is.
> We also have a guess on the computational requirements which are mind-boggling. Just simulating tiny molecular or quantum systems with total accuracy, if it's even happened, takes a ton of computation. Every sub-particle or force in the entire universe running at the same time is unimaginable computation. Plus, the Creator makes them all work exactly as desired with the same, observed patterns. We can't get cellular automata to work like we want but the Creator is directing a universe of simple interactions to cause specific behavior, global and local?
Those computational requirements exists within our universe but if we are talking about a simulation or a creator what seems like an impossible computational requirements... maybe this universe is being ran on a their version of a calculator.
One of the biggest contributors to pollinator decline is loss of habitat for native bees. Most bees are actually solitary (don’t live in hives) and live in little crannies or holes in the ground.
One of the coolest things I’ve come across recently is the idea of “bee homes” that you can put in your garden to provide habitat for bees. I’ve bought a couple beautiful wooden units from Scopa and we just got our first bee resident this week!
If you've got a lot of carpenter bees in your area, those bee homes are actually a good idea, otherwise they'll burrow into the wood of the house: https://imgur.com/a/cxQFNWG
Hey there, I actually make these homes! Unfortunately, carpenter bees don't have any interest in the model we have now as they like to make their own holes in softwoods. I'll be working with a professor that specializes in larger carpenter bees (Xylocopa) over the winter to build some habitats that are suitable for them, though, and plan to test them at her research site in the spring.
We design our blocks so that they are easy to clean. They are pre-split, but held together with UV resistant bands so that they are tightly sealed.
The need to clean them mostly comes from the agricultural sector though, and isn't required for small habitats with only a handful of nesting sites. My partner is a wild bee researcher, and they've been using these in the field for about 60 years. As long as the density is low, the requirement for cleaning them is optional.
As much as I can appreciate Doctorow’s exasperation, I find his dismissal of the doomer vs accelerationist debate rather pithy. I would love to be convinced that “dumb” LLMs can never gain sentience (or be finagled into sentience with a wrapper).
What is the actual argument for why that’s true?
(I realize you could turn the question around and ask why I think it might be possible in the first place, but I feel like my expectations have been blown out of the water so regularly and so increasingly frequently that I can’t default to being a naysayer anymore)
Just increasing the size of pre-trained LLMs is not considered a likely simple path to AI by most professionals working in the technical side of the field.
Thank you, that was a fascinating talk and I learned quite a bit.
However, it did not provide a convincing argument as to why LLMs cannot be a part of a "doomer" AI. In fact, I got the opposite vibe from Andrej explaining expected future developments. The whole section on System 2 thinking sounds like a layer constructed around dumb LLMs that would result in vastly improved and more generalizable intelligence.
I agree that just scaling the size of LLMs is probably not sufficient for AGI...but that just seems like one relatively minor piece of all the possible ways it might be achieved.
No argument from me. LLMs would be a component of AI, much like we kind of have long term read only memory. But the extra bits could in the form of some dynamic functions on each tensor network node (G* functions LOL!)
It's interesting to go back and read the MIT OpenAI story from 2020 to hear how thinking about these moral implications was an important part OpenAI.
Most professionals didn't think we were close to surpassing human capability in chess, go, or dota, until after it happened. I've seen little evidence of expert domain knowledge improving AI forecasting ability, if anything it seems the experts are often late to the party.
Besides expert consensus, is there any other actual argument against LLMs achieving generalizability?
Well there are solid technical reasons, as described in the video. One of them is based on that these models are 'pre-trained' and AGI may be a result of a more dynamic knowledge base that can change more than just the local context and update the model, as our brain does.
Andrej also suggests that an attribute of a more advanced AI would have the ability to ask it to spend longer thinking to get a better answer, like a chess engine.
This said, expert consensus is probably the best answer we have. It's not like the consensus of a bunch of youtube vids and articles that only exist for getting clicks. These experts are famously sharp. I have done his course video series (it took a huge effort, even though he is an amazing lecturer) and had existing python and linear algebra experience and I understand his argument.
>these models are 'pre-trained' and AGI may be a result of a more dynamic knowledge base
Why couldn't the knowledge base be used in conjunction with the LLM? As the GP said, why can't LLM's gain sentience or be finagled into sentience with a wrapper'. The Knowledge base you're describing is the wrapper.
>Andrej also suggests that an attribute of a more advanced AI would have the ability to ask it to spend longer thinking to get a better answer, like a chess engine.
This is another method that is already being deployed with LLMs. So the question stands, why won't LLMs be the foundation for nearing AGI?
For my money, LLMs likely are that base. AI Experts are either too shy from the memory of AI winters past to see the nose on their faces, or too busy developing paradigm breaking models to care. Regardless of what Chomsky or any other 'expert' says should be possible, the practical results of LLM growth are literally speaking for themselves.
Maybe we should have suspected a 'large language game' to be the catalyst for AGI from the start. Was human intelligence truly general before we developed language? Could it be general without it?
I don't know the actual answer to your question, but I do know that when you make beer from wild yeasts you get a really low alcohol content, like 1 or 2% ABV. There was an episode of Brew Masters^[1] that I remember watching where they collected wild yeast from a date tree in Egypt and used the yeast to make beer, and the ABV content was around 1% if I remember correctly. The brewer on that show was the head brewer at Dogfish Head, so if that's as good as he could get I doubt this company would be able to do much better. Modern alcohol is made using brewing yeast which has been selectively bred so that it has much higher alcohol tolerance, which is why we can make things like beer and wine with the alcohol percentages we're accustomed to today.
So this is pure speculation, but if I had to guess their 45M year old yeast can probably only be used to make a beer that's so watered down that it just isn't that appealing to a modern consumer.
That show was great and inspired me to get into brewing. It was a great hobby and I highly recommend it, even if you're not a drinker. Friends and family will gladly drink any excess beer you make.
I couldn't find any public statements from the brewery as to why they closed shop. A bit of snooping landed me on Raul's LinkedIn page (https://www.linkedin.com/in/raul-cano-00a4b6) where it doesn't seem to have the brewing company listed under his experience, so perhaps there was some falling out?
Sure, but it would be more constructive to explain why it's wrong. GP's comment has some interesting ideas and it would be better to directly refute them than just claiming they're "factually wrong".
> The question is really how to address the despair.
It's known that people have genetic predispositions to depression. People experience depression due to many causes. Not all causes are known and we don't have very good understanding of the mechanisms, but it's just not damn likely that we can treat this as a purely psychological / environmental issue.
Here are some "sleepers" that people have built over the years [1,2,3]
[0] https://thelaserhive.com/product-category/powermac-g4-conver...
[1] https://old.reddit.com/r/sleeperbattlestations/comments/1gva...
[2] https://old.reddit.com/r/sleeperbattlestations/comments/1871...
[3] https://old.reddit.com/r/sleeperbattlestations/comments/t7tz...