There is a comment in the intro of that book about the pinnacle of human labor being the simulation of consciousness. Very prescient for being written in the 60s.
I’ve felt the same way. It’s so inefficient to have two patterns - OLAP and OLTP - both using SQL interfaces but requiring syncing between systems. There are some physical limits at play though. OLAP will always take less processing and disk usage if the data it needs is all right next to each other (columnar storage) where as OLTP’s need for fast writes usually means row based storage is more efficient. I think the solution would be one system that stores data consistently both ways and knows when to use which method for a given query.
In a sense, OLAP is just a series of indexing strategies that takes OLTP data and formats it for particular use cases (sometimes with eventual consistency). Some of these indexing strategies in enterprises today involve building out entire bespoke platforms to extract and transform the data. Incremental view maintenance is a step in the right direction - tools like Materialize give you good performance to keep calculated data up to date, and also break out of the streaming world of only paying attention to recent data. But you need to close the loop and also be able to do massive crunchy queries on top of that. I have no doubt we'll get there, really exciting times.
Completely agree. All of the pieces are there and it's just waiting to be acted upon. I haven't seen any of the major players really doubling down on this, but would be so compelling.
Have there been any updates to Claude 3.5 Sonnet pricing? I can't find that anywhere even though Claude 3.7 Sonnet is now at the same price point. I could use 3.5 for a lot more if it's cheaper.
One thing I have not seen commented on is that ARC-AGI is a visual benchmark but LLMs are primarily text. For instance when I see one of the ARC-AGI puzzles, I have a visual representation in my brain and apply some sort of visual reasoning solve it. I can "see" in my mind's eye the solution to the puzzle. If I didn't have that capability, I don't think I could reason through words how to go about solving it - it would certainly be much more difficult.
I hypothesize that something similar is going on here. OpenAI has not published (or I have not seen) the number of reasoning tokens it took to solve these - we do know that each tasks was thoussands of dollars. If "a picture is worth a thousand words", could we make AI systems that can reason visually with much better performance?
Yeah this part is what makes the high performance even more surprising to me. The fact that LLMs are able to do so well on visual tasks (also seen with their ability to draw an image purely using textual output https://simonwillison.net/2024/Oct/25/pelicans-on-a-bicycle/) implies that not only do they actually have some "world model" but that this is in spite of the disadvantage given by having to fit a round peg in a square hole. It's like trying to map out the entire world using the orderly left-brain, without a more holistic spatial right-brain.
I wonder if anyone has experimented with having some sort of "visual" scratchpad instead of the "text-based" scratchpad that CoT uses.
A file is a stream of symbols encoded by bits according to some format. It’s pretty much 1D. It would be susprising that LLM couldn’t extract information from a file or a data stream.
I think this is a key argument in how powerful AI can become. We may be able to create incredibly intelligent systems, but at the end of the day you can’t send a computer to jail. That inherently limits the power that will be given over to AI. If an AI accidentally kills a person, the worst that could be done to it is that it is turned off, whereas the owners of the AI would be held liable.
When steam and coal engines gave way to gas and electric engines in factories, it took decades before factories were reconfigured to adjust to the smaller sized engines that didn't require one major axle running through the entire factory. As a consequence the first gas engines were huge - over time they shrunk. I bet the same will happen with robotics, where humanoid will be the primary form factor at first for general tasks, then more efficient forms will emerge as processes are updated.
These efficiënt forms are there. In use. Proven. Have been for decades.
Cranes, carts, lorries, conveyor belts (with vision), my robot vacuum cleaner, my bread baking machine, a car wash, a dishwasher, the ticket gates at the underground, the coffee machine at our office and so on.
Yet, there are still millions of humans working in factories.
There is value to the human form, our versatility and adaptability.
A machine that replicated a human would have incredible economic value (though not for the people whose jobs it replaced). A machine that exceeded a human in versatility, e.g. by having more arms, even more so.
Yes, there are. But that's not caused by lack of humanoid robots, but because these humans are rediculously cheap.
There's absolutely no way that automation and tech can undercut (effective) slave labor.Though already this is happening: your tshirts and socks are probably not hand-knitted by "Bangladeshi children" or Chinese "prisoners", but by machines tailor-made (pun intended) for t-shirt and sock-knitting. This is happening slowly, and in the supply chain there's still loads of manual labor - cotton picked and processed for these socks is more and more automated, but still requires armies of cheap farm-hands.
And then it's really hard to undercut the price of cheap "western" labor with full automation in many sectors. Part of that is due to some form of 90/10 rule: the last 10% of automation is magnitudes more expensive. And many automation leads to a higher TCO, as the people programming, maintaining and optimizing the robots are far more expensive than an "army" of low-wage workers.
Humanoid robots may be an answer in some future, but currently and in near future will certainly not solve these economics. If ever.
I wouldn't be surprised if the slowest part of the system is the API call to a legacy warehouse management system that takes several seconds to respond to get the next bin to target.
I've felt this way and have started celebrating inputs - the work involved in building something - rather than outputs (the sale, the deployment, etc.). Once it closes or the deployment is successful it's always onto the next thing immediately and there's no time to sit back and reflect on the hard work and enjoy the time spent with others in the process.
I think the compelling difference is truthfulness. There are certain people / organizations that I trust their synthesis of information. For LLMs, I can either use what they give me in low impact situations or I have to filter the output with what I know as true or can test.
After a recent 3 week driving trip through Europe, I can anecdotally back this up. It’s not even the manual transmission, but also the much smaller roads with no shoulder where at times you meet a car, have to slam on the brakes, and decide in the moment whether you or the other car will back up to a turn out. You have to pay constant attention and there’s little room for looking at phones, eating, etc. In the US we have such large roads and shoulders that you can zone out and are more easily tempted to take your eyes off the road.
This is backed by infrastructure studies. It's called traffic calming and it's done on purpose in places with advanced infrastructure such as the Netherlands.
In the same vein, I encourage neighbors to park their cars on the street rather than in their garage or driveway so the neighborhood street becomes narrower for traffic due to a row of parked cars on each side of the street. Works better than posted speed limit.
Demonstrably incorrect on my street, which has hotly contested on-street parking that narrows the roadway to a single lane. Cars have to stop and reverse into parking spaces to pass each other. Yet children play ball games in the street, and stop to move out of the way of the cars, with the cars barely needing to brake.
Children are not mindless suicide machines; they learn at a tremendous rate from their environment. The children you see around you might not survive on my street, but that's because they haven't lived there for all of their lives.
