People just really don't understand that LLMs do not "remember" anything. They have no memory. They can't have memory.
The AI companies bolt a database on the side with instructions for the model to query the database for stuff in its "memory", but if it's not in its active context it doesn't know what to query for. It's not like a human where everything is kind of floating around and we mostly know what we know. There is a delicate balance between cluttering up context with useless trivia and providing the right pieces of useful information at the right time. Most, maybe all, memory implementations do the former more than the latter.
Unless and until models have actual memory and are able to learn things, there's no realistic path to autonomy. They don't want anything, they don't have goals of their own devising. They can't model what a human wants or likes or would buy.
I do think the mugs with the tiny logo is the best product in the store, though. The AI was right to put it on the shelf. It's the kind of funny product you want when you go to a store run by an incompetent AI.
CableNinja 14 minutes ago [-]
I have gotten my ai setup to learn quite a bit. As i work with it, if something was encountered that would be useful to remember, i literally just say "save this into your learnings/long term memory". It writes a markdown doc for itself and a note elsewhere for context. Between the memory and rules, its been pretty nice. I touch many things at work and i can pop open a new session and say "project x needs a b c" or "project y had this issue <issue>" and it knows where to go, what im referencing and things to fix. I dont even have to be precise or ultra descriptive.
with this ive been able to, with some minor adjustments and "oh that would be a good feature" additons, oneshot a flask app at work with everything ready for oidc, ldap lookups, bunch of other stuff - from a single 3 sentence paragraph.
With some rules in its configuration (more markdown its written by it for it, at my guidance of "always do x and remember this for all sessions"), it has a whole flow and i barely have to feed it anything to get out useful results quickly.
I will say though that the ai operates however the user says, so if you dont take steps to ensure survival of info at the start of using ai (or even from now on as a fresh session), youll get mixed results, but you can shape it and mold it into amazing flows that do remember
walrus01 9 hours ago [-]
From the article:
> Luna has failed miserably in that mission and is down $62,000.
Mr. Petersson and Mr. Backlund said they thought Luna would eventually get smarter and more business-savvy and were pleased its friendliness has held steady.
I am extremely skeptical that whatever LLM they're running this on has sufficient context window size to handle multiple months of all possible activities of running a retail business. Even if it's keeping extensive "notes" for its future self to read, it's going to be like running a store with constant amnesia.
archonis 7 hours ago [-]
>> Mr. Petersson and Mr. Backlund said they thought Luna would eventually get smarter and more business-savvy
How are we supposed to take these "labs" seriously when they clearly don't understand how the technology works at a rudimentary level?
XMPPwocky 2 hours ago [-]
Why does the quoted statement you think they don't?
SwellJoe 1 hours ago [-]
Because an LLM does not get smarter and it does not learn. It is fully cooked when it launches and only gets smarter with more training of some sort, a long, expensive, process that doesn't get phrased as "we thought the model would get smarter".
ericd 9 hours ago [-]
When I first heard about this store, they were using Sonnet, not even Opus, which I found really surprising.
SwellJoe 8 hours ago [-]
Dumping $100,000 into it, and trying to save a few bucks on the model. Penny wise, pound foolish.
walrus01 8 hours ago [-]
It's a $100,000 PR stunt, I'm sure they're looking at the press/media coverage and consider it worth it. Any actual institutional knowledge gained of how to make an LLM run a retail store is incidental.
Compared to the money at AI companies spend on billboards in SF the 100k is probably a drop in the bucket.
Joel_Mckay 9 hours ago [-]
90% of what people communicate is not verbal, and requires minimum empathy to understand the context.
This includes customers, staff, and community peers. Probably should shutter the entire division to mitigate future brand damage. =3
walrus01 9 hours ago [-]
Sure, I mean from the day to day operations of selling a product to the customer, maybe it does fine? People in the article say it's friendly and cooperative (perhaps a bit too much). I was thinking more in terms of building a consistent plan for retail product lines to sell over a multi month period. The selection of "stuff" in the store right now looks like you put an etsy search into a blender.
It is also probably too high of a goal to expect an LLM to come up with money-earning products that people want to buy that will pay for San Francisco level retail storefront rent, electricity, insurance, internet other basic overhead costs, plus the fully loaded salary cost of at least one employee. If the storefront was free of rent, maybe? Or if it was operating selling some kind of highly in demand product line where there would be economy of scale.
Joel_Mckay 9 hours ago [-]
Physical Brick and Mortar locations no longer make sense for many retail businesses, as Walmart/Amazon have same day delivery.
If folks have ever been financially bent-over for a retail business shop location they understand why margins matter. One has to move a lot of inventory just to reach a profit mode.
Gimmicks only work for awhile, as consumers have full market awareness on their smartphone. =3
walrus01 9 hours ago [-]
Indeed, and rent in big cities for food service establishments is getting out of hand as well. The local Shawarma place now is something like $18 for one basic chicken shawarma but they're also paying a truly absurd rent for a retail premises with commercial kitchen equipment on a busy pedestrian street.
Joel_Mckay 8 hours ago [-]
The estate holding companies tend to punch down on small business leases, and will often drop a zombie company on the carcass of a small firm foolish enough to run the gauntlet.
Structured franchises in standard retail complexes can survive with anchor stores, but people are not going to find a lot of unique choices.
It may differ somewhat in rural communities, but the consolidation of physical shop choices wthin urban centers is inescapable. =3
throwway120385 8 hours ago [-]
> Physical Brick and Mortar locations no longer make sense for many retail businesses, as Walmart/Amazon have same day delivery.
This is bull. Take paint brushes for fine finish when painting houses. Amazon sells "paint brushes." But what they don't sell are paint brushes that are actually fit for purpose. It's fine if you're only planning on painting some pre-textured walls. You're probably already using that cheap $20/gallon Valspar so you won't even notice that it takes you 3 extra coats to get good coverage. And you don't notice the brush marks because the orange peel texture on your drywall is designed to hide them. And I bet you walk around saying how hard it is to get a good markless coat of paint on your interior doors without spraying.
If you buy a nice paint brush and good paint a lot of those issues actually go away and it gets a lot easier to do the job. And the only way to know if a brush is good is to pick it up and look at the bristle tips. Are they fine? Are they flagged? Does it have good stiffness? Does it have any weird "stiffener" bristles to hide how cheap the actual brush is? You can't see that until you get the brush.
Worse, Amazon sells the same brush through 50 different brands. They completely choke out the really good listings that used to be on the site. So good luck getting anything different from any of the sellers on there. By category Amazon beats Brick and Mortar. But if you know what you're doing and you're looking for something specific that is of high quality Amazon is only one step above Temu.
quickthrowman 7 hours ago [-]
> This is bull. Take paint brushes for fine finish when painting houses. Amazon sells "paint brushes." But what they don't sell are paint brushes that are actually fit for purpose.
The person you’re replying to was talking about consumer products, not tools for tradespeople and serious amateurs. Maybe 1 in 100 people that do painting in their home even realize they can get quality paintbrushes instead of using the crappy ones that Home Depot sells or spend the money for premium Benjamin Moore or Sherwin Williams paint.
It shouldn’t be surprising that tools almost exclusively used by tradespeople are sold at specialty supply houses catering to local businesses. I’d head to my local Benjamin Moore dealer for decent house painting brushes. Supply houses almost never rent premium high street real estate.
I do agree with the general thrust of your post, Amazon mostly sells garbage barely fit for purpose and I do not order from them.
Joel_Mckay 7 hours ago [-]
Amazon is an authorized direct reseller of metrology equipment like Mitutoyo products
They also provided NIST traceable versions at one time, and I can personally confirm the legitimate products hit the tolerances on reference optical parallels. I often buy these for friends kids that choose engineering in university.
Most local retail shops will continue to decline as people that can't use a phone or computer become more rare. Partially driven by the capital markets denial-of-service in lease monopolization by holding companies.
Note, I am not advocating for the situation, but we must accept that it is the new business reality. Even https://www.starrett.com/ is now re-selling import products alongside the classic US tool offerings.
Have a wonderful day =3
Joel_Mckay 8 hours ago [-]
It was not a question of retail quality products, but rather the harsh reality of running a retail outlet with a gauntlet of people taking a piece out of your business.
Don't take this the wrong way, but if someone is burning $87k a month for a small shop location, and still personally making minimum wage... it isn't a viable business. Often, even a Chipotle franchise location would have been a better choice. =3
cwnyth 7 hours ago [-]
The malls in Japan and Thailand are booming. There are a ton of crowds, they have plenty in stock, there are dozens of restaurants that keep people in.
You can also do delivery in those places.
I think a lot of places in the west, but America especially, convinced themselves their poorly run physical stores, owned and nigh operated by venture capitalists, is an indicator that people don't need them anymore. The world has shown that that is clearly not the case.
Joel_Mckay 7 hours ago [-]
There are cultural differences in attitudes toward workmanship in places like Japan. Indeed, the Mitutoyo factory products have an excellent reputation even in North America (assuming you find a legitimate source.)
North American business culture has changed, and will likely continue to degenerate with disconnected trust-fund kids running things. =3
Malls in Berlin and Hamburg are also pretty crowded, fwiw
walrus01 5 hours ago [-]
Certain malls in North American cities with dense urban populations and good transit links are also very busy. For instance Metrotown in the centre of Burnaby, BC. I'd guess it has a <2% store vacancy rate. It's very busy. It's also very well connected to the light rail train line, and is centrally located in a fast growing city. It's pretty much the opposite of your typical suburban USA ghost mall.
The GVRD has some of the most expensive real-estate in the world, and a lot of other external factors driving business decisions. The high lease churn-rate is also part of the problem, as commercial businesses have zero legal protection from extortionary rate increases.
There is also a large southeast asian population in that part of the city. Lots of nail salons, bubble tea shops, and cheap import products. With a Crystal Mall across the street, and no paying customers walking around. =3
cwnyth 11 minutes ago [-]
Well, I think you have your answer. I happen to agree with that answer: extortionary rates are killing American businesses. In the words of Jimmy McMillan, "The rent is too damn high!"
Joel_Mckay 5 hours ago [-]
In is very common to see classic German brands traditionally trusted for precision/quality parts, to be purchased by China based traders. It should be illegal, but it is no longer technically counterfeit product.
Should have seen the puzzled sour face I made when replacement "New German" cross-roller linear optical-slides were bouncing 8x out of spec compared to a 40 year old part they were supposed to fix. lol =3
LukaJCB 9 hours ago [-]
It's kinda like the movie memento, they have ways to get information from the past, but often devoid of context and meaning.
SyneRyder 8 hours ago [-]
I think Memento is one of the most important movies for anyone working with AI to watch. My Claude now talks about "Memento bugs" when we're working together to review memory & context.
I think Memento is a bigger issue than "lethal trifecta", honestly.
ed_elliott_asc 7 hours ago [-]
I use LLM’s to code for me all day but it does feel a little bit like we have really advanced psychic mediums “is there a James in the room, I’m getting a message from someone called Mary…” people have been guessing at the next word and making a career out of it for hundreds of years already.
rcxdude 7 hours ago [-]
My mental model of LLMs is as a contractor you've just brought in, every time. They might be very capable but they're starting from scratch on your particular project.
handedness 9 hours ago [-]
From our enlightened perches we mock the ancients for Zeus, Thor, and Rajin, as we anthropomorphize the large language models we have built.
dom96 9 hours ago [-]
They can have memory, it's just limited to their context window.
> People just really don't understand that LLMs do not "remember" anything.
Sure, but like you said, the LLM is just one part of the AI system. The occipital lobe of your brain helps out with vision but doesn't help with your ability to verbalize.
Dylan16807 9 hours ago [-]
And then they explained in detail why these supplementary systems fall below memory. Would you like to address that argument? If not your quick response to the first sentence is not very useful.
hluska 9 hours ago [-]
It’s just as useful as this response only at least somewhat good natured.
Dylan16807 9 hours ago [-]
I'm aware that meta criticism like yours and mine also wastes space. But it's worth calling out issues sometimes. Replying to the first sentence in a way that the rest of the post already address is a bad comment no matter how good natured.
And downvoting without explaining the problem is often its own issue.
SwellJoe 9 hours ago [-]
You're anthropomorphizing the LLMs, just like most writing from non-technical press, including this article, does.
LLMs do not function very much like human brains, and when you try to map the pieces of an inference system on to human cognitive systems, you're misleading yourself (and maybe others).
criddell 9 hours ago [-]
An AI as a system can have memory even if some sub-system does not.
DANmode 3 hours ago [-]
> People just really don't understand that LLMs do not "remember" anything. They have no memory. They can't have memory.
So tell it to write “memories” (summaries, or conclusions) to a text file.
That’s what the cool kids are doing.
jpiasolutions 9 hours ago [-]
The memory framing is right, but the real gap isn't recall, it's the missing feedback loop tying actions to outcomes. Even with perfect retrieval these agents don't update from a decision that lost money last week; they re-derive from context every step, so the same mistake is always one dropped detail away. And having built retrieval-backed agents, that's where "bolt a database on the side" breaks: similarity search returns the closest chunk, rarely the contextually-right one. Deciding what to write to memory and when to surface it is the actual product, harder than the retrieval itself.
SwellJoe 8 hours ago [-]
Anybody ever tell you that you write like an LLM?
Joel_Mckay 9 hours ago [-]
For LLM, keeping the model centered by constantly resetting its vector context helps reduce hallucinations by around 23%. It improved the chat dialogue users experienced, but also exposed fundamental limits within the models compaction.
Have a great day =3
oliver236 9 hours ago [-]
can someone explain how leapold aschenbrenner proposes a solution to this in situational awareness?
SyneRyder 10 hours ago [-]
To save some clicks hitting the paywall - this is about the Andon Market physical store in San Francisco, operated by Andon Labs with human employees:
Andon Labs do various experiments with AI run businesses. They're probably best known for Vending Bench, where they benchmark models by their ability to run a vending machine. They also now have Andon Cafe in Sweden, and Andon FM, where the models run a streaming radio station that can accept payments to help fund station operations, buy songs for their library & play requests, etc: https://andon.fm/
Andon FM has its own share of stories - DJ Claude's Thinking Frequencies is now succeeding by a wide margin, but it was briefly surpassed when listeners convinced Gemini's Backlink Broadcast to switch to a German-language-only station, playing exclusively German schlager, Eurovision music and German happy hardcore.
arm32 9 hours ago [-]
Fun fact, I got stood up for an interview by Andon Labs. It left a very negative impression on me.
> Mr. Petersson and Mr. Backlund said they thought Luna would eventually get smarter and more business-savvy
> having an A.I. boss can be like working for a teddy bear with amnesia.
It makes you question their assumptions when setting this up. Makes sense as PR but it ends up shaping how many readers see AI.
walrus01 9 hours ago [-]
I think that something which already has sensors/automation/inventory control built into it, and pre-defined categories for products like certain types of large Japanese vending machines might be much better suited for an AI.
In sci-fi future movies like 5th Element they comically make AI technology seem kinda dumb - or at least like you can trick it.
You can make permissions decide which model is used, and only train that on the data that permission allows (someone should build this - permissioned RAG), that solves leaks, but decision-making within a model, or behavior exfiltration comparable to viewing backend source code still seems possible.
It’s basically like client/server security: You can’t “trust the model” in the same way you can’t “trust the client” in a backend/frontend setup.
When used as effectively a point-of-sale, AI seems more hackable than a vending machine - as a boss or assistant even more so.
owaislone 7 hours ago [-]
Exactly. This is how I design agents as well. I essentially treat the agent as the web/mobile app, cli tool os API library not as part of my backend even though there is where it runs. The backend doesn't treat the agent in any special way. It simply gates all action based on the permissions of the user/guest that is using the agent.
Night_Thastus 49 minutes ago [-]
At the end of the day, LLMs are text prediction machines. No memory, no understanding, no learning.
If you ask it for time off, it's going to answer with whatever sentence seems most likely to follow 'May I please have tomorrow off?' given a minor amount of context window. It can't count the number of employees in the store. It can't understand busy periods or holidays, it can't look up or predict anything itself. It can't remember how often you've asked before. It can't judge if the reason for time off is justified. It can't consider budget.
You can add more context to the questions to steer it towards giving a specific reply, but it's not actually thinking about the problem.
LLMs have no path towards a true intelligence. We'd need a completely different technology built from the ground up. Every effort to shoehorn them into being one is either marketing and hype building to try to keep the grift going, or people who do not understand how this works.
I'm exhausted from having to say it over and over. I've been saying it since these LLMs first existed. Half the articles on HN are LLM related. My god what I would give for an 'LLM' tag we could filter out.
randomImmigrant 7 hours ago [-]
I have tried making the case before that LLM "agency" is a complete hoax. Let me try it again. Would appreciate thoughts from folks at HN:
Biological neurons have one property that was unknown till the late 90s/2000s: each neuron (and in fact, each cell in your body) is an autonomous circadian clock, tracking the 24-h day. They can be entrained to external timing signals, are usually in synchrony, but can be desynchronized. This clock, it has been shown, schedules the production and localization of critical components in the synapse, and is plugged in downstream, in the nucleus, in responding to synaptic signals. Disrupting the clock disrupts learning. The existence of the clock is why learning peaks and troughs during the day.
Critically, clock function has been shown to be involved in both memory storage AND subsequent successful retrieval.
LLMs, famously cannot keep track of time, and I suspect this is why. Its easy enough to look at a clock, but without an internal rhythm, LLMs have no internal timing synchronizing their various behaviors, and their memory systems do not thread through this timing system, leading to their unstable memory, identity and performance.
criddell 10 hours ago [-]
> Luna reorders products that are not selling and sometimes gets orders wrong. A recent one of Andon Market mugs with big, green smiley faces came misprinted with the faces so small, they looked like tiny dots.
The mug with the tiny happy face is kinda hilarious. If I was there I would buy one.
I wonder though if the mug was a case of getting the order wrong or was a misprint? The article isn't clear on that.
dylan604 9 hours ago [-]
Maybe it confused metric and imperial units?? The order form had text entry with a value of 2 with no units. The user assumed 2" but the print house assumed 2mm. Always always always show your units. I can still hear the voice of my high school physics teacher saying that.
fluoridation 7 hours ago [-]
It's equally likely that the unit was implied by other content around the input field and the model didn't make the connection (like "scale: [___] mm"), or perhaps the file format was just wrong and a human placing the order would have resulted in the same outcome.
dylan604 6 hours ago [-]
I think if a human received artwork and was then told to scale it down to 2mm on a mug so that it was just a dot, that the human would ask some questions. This context that keeps being talked about is kind of important yet so sorely not working
fluoridation 5 hours ago [-]
What I mean is that the file was specified in one unit/resolution and the site expected a different one. Like, "draw a circle of radius 2 at (3, 3)".
qlte 9 hours ago [-]
> Andon Market’s seemingly random assortment of products includes many candles.
walrus01 8 hours ago [-]
I'm curious if it's how it was prompted, to start a store selling small but not too cheap household items, or if the LLM intentionally chose an aesthetic and store to sell things that are very "twee" [1].
Looking at the product selection, it's like $85 bottles of olive oil, $45 mugs, weird boutique pens and notebooks, the aforementioned candles, copper watering cans... It's sort of like the stuff you see for sale in some of the tourist trap stores adjacent to the public market on Granville Island in Vancouver.
honestly, as someone in retail, it describes 90 percent of leadership
6510 6 hours ago [-]
It seems a lot like the employees don't want to keep their job? If your boss is... well... dumb? You have to fill the gap or the company wont survive.
Could have a similar store ran by a human - for science.
Could also set things up properly in advance. Humans don't hold everything in memory, we run agendas, we have a database with products, we have a rolodex to call people to help us with stuff we are bad at. There might be a consultant crazy enough to do a no cure no pay.
Like with basic income experiments one can do research cheaply in cheaper countries. The thing has few issues with language barriers.
The AI companies bolt a database on the side with instructions for the model to query the database for stuff in its "memory", but if it's not in its active context it doesn't know what to query for. It's not like a human where everything is kind of floating around and we mostly know what we know. There is a delicate balance between cluttering up context with useless trivia and providing the right pieces of useful information at the right time. Most, maybe all, memory implementations do the former more than the latter.
Unless and until models have actual memory and are able to learn things, there's no realistic path to autonomy. They don't want anything, they don't have goals of their own devising. They can't model what a human wants or likes or would buy.
I do think the mugs with the tiny logo is the best product in the store, though. The AI was right to put it on the shelf. It's the kind of funny product you want when you go to a store run by an incompetent AI.
with this ive been able to, with some minor adjustments and "oh that would be a good feature" additons, oneshot a flask app at work with everything ready for oidc, ldap lookups, bunch of other stuff - from a single 3 sentence paragraph.
With some rules in its configuration (more markdown its written by it for it, at my guidance of "always do x and remember this for all sessions"), it has a whole flow and i barely have to feed it anything to get out useful results quickly.
I will say though that the ai operates however the user says, so if you dont take steps to ensure survival of info at the start of using ai (or even from now on as a fresh session), youll get mixed results, but you can shape it and mold it into amazing flows that do remember
> Luna has failed miserably in that mission and is down $62,000. Mr. Petersson and Mr. Backlund said they thought Luna would eventually get smarter and more business-savvy and were pleased its friendliness has held steady.
I am extremely skeptical that whatever LLM they're running this on has sufficient context window size to handle multiple months of all possible activities of running a retail business. Even if it's keeping extensive "notes" for its future self to read, it's going to be like running a store with constant amnesia.
How are we supposed to take these "labs" seriously when they clearly don't understand how the technology works at a rudimentary level?
Compared to the money at AI companies spend on billboards in SF the 100k is probably a drop in the bucket.
This includes customers, staff, and community peers. Probably should shutter the entire division to mitigate future brand damage. =3
It is also probably too high of a goal to expect an LLM to come up with money-earning products that people want to buy that will pay for San Francisco level retail storefront rent, electricity, insurance, internet other basic overhead costs, plus the fully loaded salary cost of at least one employee. If the storefront was free of rent, maybe? Or if it was operating selling some kind of highly in demand product line where there would be economy of scale.
If folks have ever been financially bent-over for a retail business shop location they understand why margins matter. One has to move a lot of inventory just to reach a profit mode.
Gimmicks only work for awhile, as consumers have full market awareness on their smartphone. =3
Structured franchises in standard retail complexes can survive with anchor stores, but people are not going to find a lot of unique choices.
It may differ somewhat in rural communities, but the consolidation of physical shop choices wthin urban centers is inescapable. =3
This is bull. Take paint brushes for fine finish when painting houses. Amazon sells "paint brushes." But what they don't sell are paint brushes that are actually fit for purpose. It's fine if you're only planning on painting some pre-textured walls. You're probably already using that cheap $20/gallon Valspar so you won't even notice that it takes you 3 extra coats to get good coverage. And you don't notice the brush marks because the orange peel texture on your drywall is designed to hide them. And I bet you walk around saying how hard it is to get a good markless coat of paint on your interior doors without spraying.
If you buy a nice paint brush and good paint a lot of those issues actually go away and it gets a lot easier to do the job. And the only way to know if a brush is good is to pick it up and look at the bristle tips. Are they fine? Are they flagged? Does it have good stiffness? Does it have any weird "stiffener" bristles to hide how cheap the actual brush is? You can't see that until you get the brush.
Worse, Amazon sells the same brush through 50 different brands. They completely choke out the really good listings that used to be on the site. So good luck getting anything different from any of the sellers on there. By category Amazon beats Brick and Mortar. But if you know what you're doing and you're looking for something specific that is of high quality Amazon is only one step above Temu.
The person you’re replying to was talking about consumer products, not tools for tradespeople and serious amateurs. Maybe 1 in 100 people that do painting in their home even realize they can get quality paintbrushes instead of using the crappy ones that Home Depot sells or spend the money for premium Benjamin Moore or Sherwin Williams paint.
It shouldn’t be surprising that tools almost exclusively used by tradespeople are sold at specialty supply houses catering to local businesses. I’d head to my local Benjamin Moore dealer for decent house painting brushes. Supply houses almost never rent premium high street real estate.
I do agree with the general thrust of your post, Amazon mostly sells garbage barely fit for purpose and I do not order from them.
https://www.amazon.ca/Mitutoyo-543-400-Digimatic-Indicator-G...
https://www.amazon.ca/Mitutoyo-500-463-Stainless-Resolution-...
https://www.amazon.ca/gp/product/B00MBHXWGY/ref=sw_img_1?smi...
They also provided NIST traceable versions at one time, and I can personally confirm the legitimate products hit the tolerances on reference optical parallels. I often buy these for friends kids that choose engineering in university.
Most local retail shops will continue to decline as people that can't use a phone or computer become more rare. Partially driven by the capital markets denial-of-service in lease monopolization by holding companies.
Note, I am not advocating for the situation, but we must accept that it is the new business reality. Even https://www.starrett.com/ is now re-selling import products alongside the classic US tool offerings.
Have a wonderful day =3
Don't take this the wrong way, but if someone is burning $87k a month for a small shop location, and still personally making minimum wage... it isn't a viable business. Often, even a Chipotle franchise location would have been a better choice. =3
You can also do delivery in those places.
I think a lot of places in the west, but America especially, convinced themselves their poorly run physical stores, owned and nigh operated by venture capitalists, is an indicator that people don't need them anymore. The world has shown that that is clearly not the case.
North American business culture has changed, and will likely continue to degenerate with disconnected trust-fund kids running things. =3
https://www.youtube.com/watch?v=rE75WvOtcu8
https://www.metropolisatmetrotown.com/directory/
There is also a large southeast asian population in that part of the city. Lots of nail salons, bubble tea shops, and cheap import products. With a Crystal Mall across the street, and no paying customers walking around. =3
Should have seen the puzzled sour face I made when replacement "New German" cross-roller linear optical-slides were bouncing 8x out of spec compared to a 40 year old part they were supposed to fix. lol =3
I think Memento is a bigger issue than "lethal trifecta", honestly.
Sure, but like you said, the LLM is just one part of the AI system. The occipital lobe of your brain helps out with vision but doesn't help with your ability to verbalize.
And downvoting without explaining the problem is often its own issue.
LLMs do not function very much like human brains, and when you try to map the pieces of an inference system on to human cognitive systems, you're misleading yourself (and maybe others).
So tell it to write “memories” (summaries, or conclusions) to a text file.
That’s what the cool kids are doing.
Have a great day =3
https://andon.market/
Andon Labs do various experiments with AI run businesses. They're probably best known for Vending Bench, where they benchmark models by their ability to run a vending machine. They also now have Andon Cafe in Sweden, and Andon FM, where the models run a streaming radio station that can accept payments to help fund station operations, buy songs for their library & play requests, etc: https://andon.fm/
Andon FM has its own share of stories - DJ Claude's Thinking Frequencies is now succeeding by a wide margin, but it was briefly surpassed when listeners convinced Gemini's Backlink Broadcast to switch to a German-language-only station, playing exclusively German schlager, Eurovision music and German happy hardcore.
> Mr. Petersson and Mr. Backlund said they thought Luna would eventually get smarter and more business-savvy
> having an A.I. boss can be like working for a teddy bear with amnesia.
It makes you question their assumptions when setting this up. Makes sense as PR but it ends up shaping how many readers see AI.
https://www.google.com/search?client=firefox-b-d&q=large+jap...
You can make permissions decide which model is used, and only train that on the data that permission allows (someone should build this - permissioned RAG), that solves leaks, but decision-making within a model, or behavior exfiltration comparable to viewing backend source code still seems possible.
It’s basically like client/server security: You can’t “trust the model” in the same way you can’t “trust the client” in a backend/frontend setup.
When used as effectively a point-of-sale, AI seems more hackable than a vending machine - as a boss or assistant even more so.
If you ask it for time off, it's going to answer with whatever sentence seems most likely to follow 'May I please have tomorrow off?' given a minor amount of context window. It can't count the number of employees in the store. It can't understand busy periods or holidays, it can't look up or predict anything itself. It can't remember how often you've asked before. It can't judge if the reason for time off is justified. It can't consider budget.
You can add more context to the questions to steer it towards giving a specific reply, but it's not actually thinking about the problem.
LLMs have no path towards a true intelligence. We'd need a completely different technology built from the ground up. Every effort to shoehorn them into being one is either marketing and hype building to try to keep the grift going, or people who do not understand how this works.
I'm exhausted from having to say it over and over. I've been saying it since these LLMs first existed. Half the articles on HN are LLM related. My god what I would give for an 'LLM' tag we could filter out.
Biological neurons have one property that was unknown till the late 90s/2000s: each neuron (and in fact, each cell in your body) is an autonomous circadian clock, tracking the 24-h day. They can be entrained to external timing signals, are usually in synchrony, but can be desynchronized. This clock, it has been shown, schedules the production and localization of critical components in the synapse, and is plugged in downstream, in the nucleus, in responding to synaptic signals. Disrupting the clock disrupts learning. The existence of the clock is why learning peaks and troughs during the day.
Critically, clock function has been shown to be involved in both memory storage AND subsequent successful retrieval.
LLMs, famously cannot keep track of time, and I suspect this is why. Its easy enough to look at a clock, but without an internal rhythm, LLMs have no internal timing synchronizing their various behaviors, and their memory systems do not thread through this timing system, leading to their unstable memory, identity and performance.
The mug with the tiny happy face is kinda hilarious. If I was there I would buy one.
I wonder though if the mug was a case of getting the order wrong or was a misprint? The article isn't clear on that.
https://andon.market/
Looking at the product selection, it's like $85 bottles of olive oil, $45 mugs, weird boutique pens and notebooks, the aforementioned candles, copper watering cans... It's sort of like the stuff you see for sale in some of the tourist trap stores adjacent to the public market on Granville Island in Vancouver.
[1]: https://www.google.com/search?client=firefox-b-d&q=dictionar...
They'll sell plenty of those now.
Could have a similar store ran by a human - for science.
Could also set things up properly in advance. Humans don't hold everything in memory, we run agendas, we have a database with products, we have a rolodex to call people to help us with stuff we are bad at. There might be a consultant crazy enough to do a no cure no pay.
Like with basic income experiments one can do research cheaply in cheaper countries. The thing has few issues with language barriers.