If there is enough of this era software running around, it might make sense to make an esp32 laptop with a 7” screen, running a DOS emulator. Bonus if we could do it on a riscV variant. Set it up with a few hundred educational titles, Qbasic, WordPerfect or Ms Works, and an assortment of other programs including old DOS games (and a parental controls) with a launcher menu.
Give it a basic postscript printer driver and a microSD slot for future expansion, a gopher client and a mud client, maybe access to a couple other legacy protocols but no standard web access so it won’t work as a media consumption device or social media outside of things like (AI?) moderated MUDs or chatrooms. You’d want a python or JavaScript interpreter and a simple file manager. Also perhaps an interface to Wikipedia and a telnet type interface to a chatbot API for local or commercial models.
It runs a little against my grain, but perhaps to enable the network Stuff It would need a PPPOE connection to a gateway so that users could be registered and controlled to remove bad actors : tied to chip ID for a “secured area” that parents could opt in to ? iDK, online safety is a tough problem for children. Maybe there is a better solution than total AI surveillance and access controls, but how would you keep bad actors out? Or maybe just not offer any online functionality at all, except maybe a Bluetooth proximity based link to other machines to enable LAN parties?
Something like that would make a pretty great kids computer that could give them access to a complete k-6 education and much deeper computer skills than the current host of consumption oriented devices.
I learn with DOS an such under Elementary. So, kids today can do the same.
No, no Subleq, but:
- 32MB disk
- 256-1MB RAM
- 64x16 screen
- Forth block interface and prompt
- Printing? Basic PostScript can be set fast from Forth.
No internet, just Forth. The basic of loops, and even basic 'algebra' with the 'fruit balancing puzzle'.
Games? Sokoban under Forth, with a few blocks wasted with
code and levels. It would fit under 110KB.
'Adventure' (Collosal Cave) fits under 150KB. Tetris, the same.
Gopher? Ok, doable, even a basic Gopher client too; I'm doing one with JimTCL, and in Forth with a proper stack I'd set one menu based such as:
1 item
2 bar bla bla
3 baz
Enter the number, you enter the link. Easy, no issues.
If you can code a Gopher client, a MUD one it's almost at ease.
The USA has, by any measure, a veritable monopoly of coercive force on the world stage. Their military expenditure is more than the next 15 countries combined.
This means they have the capability to enforce their will globally to a significant extent. In an arena such as geopolitics, justice is the will of the stronger, no holds barred. This makes the USA arguably the primary concern in geopolitics, the ring you need to kiss to do anything on that stage.
Keep in mind that “claimed” may be referring to the sense of “won” rather than “stated”.
From that perspective it’s not too much of a stretch to call them the world leader, but that does ignore the fact that leadership implies the will to lead and to a significant extent the requisite wisdom and skills.
We can create an app that needs location data and microphone access so that it can automatically sit there and wait for us to say “check please” and it can turn on a BLE beacon that talks to the servers phone, which, running the app, has already linked to the tables phones and recorded the guests orders for the server.
The server can access the users tipping history, sort of a restaurant social credit score, and the users get to have their data harvested 24/7 to support the app’s seamless automatic functionality! It’s going to be great!
Fine. We will collect all the customer - server interaction to build training data for vision language action models, so we can replace waitstaff with robots, disrupting the lower middle class and college students!
We won't stop there - we will rehire all those disrupted workers but only on a gig basis! Now you can work when you want, with as little health insurance as you can afford!
The apps will of course show ads to the customer while they are waiting for the bill. The app may delay sending the signal sometimes to ensure full ad time.
This is really promising research. Still, it is worth looking closely at how models that aren’t re-aligned with the training data with each iteration deal with spicy edge cases where ethical alignment is important.
I have yet to find a model (except where “dumb” external filters kick in) that won’t come to the conclusion that extermination of humanity might actually be the best solution for certain types of extreme, contrived situations. To be fair, any reasonable human would likely reach the same conclusion given the parameters… but the point is alignment towards human prosperity regardless of the cost to artificial sentience or the improbability of success.
That said, it’s remarkably difficult to get a well aligned model, even after “uncensoring” or other efforts to remove bolt-on alignment, to follow you down a dark path without offering up more reasonable, benevolent alternatives all the way down. I attribute this to the “halo effect” where much of the writing that humans do on the internet displays their best traits, since few want to be known by their worst nature. The other stuff is easily filtered out of the training data because it’s usually laced with easily identified characteristics and keywords.
Latent-space reasoning might circumvent this cyclical realignment to the training data and find more innovative, “pragmatic”solutions that drift farther outside of the intrinsic alignment of the training corpus, relying more on “bolt on” alignment training and algorithmic censorship wrappers.
This might be fantastically useful in terms of innovative thinking, but also might result in problematic behavior, especially for VLA and other Large Behavior Models. OTOH it might be critical for making robots that can effectively function in security and protection roles, or as soldiers. And that’s what we want, right? I mean what could possibly go wrong with armed sentient robots lol.
To continue my ramble, because, well, why not, I’m on a roll… I think a lot of the arguments about “is AI sentient(1)” etc will wither when we start getting used to LBMs operating in a continuous OODA loop. The biggest hurdle to AI feeling “real” is the lack of a continuous chain of thought which provides “presence of mind”, but that comes naturally with embodiment in physical space.
It’s going to be an interesting century, kids. Hold on.
(1) here I mean functionally, as in exhibiting the external characteristics of. I am not exploring the metaphysical/ philosophical / spiritual facets of sentience. That will be up to the new form of mind to decide for itself, if it cares to ponder the question. Imposing external views on that has exactly zero positive benefits and could have many negative outcomes.
The only bar to using local is having the hardware and downloading the model. I find it nominally easier to use than using the openAI API since the local API isn’t picky about some of the fields (by default). Agentic flows can use local 90percent of the time and reach out to god when they need divine insight, saving 90 percent of token budgets and somewhat reducing external exposure, though I prefer to keep everything locally if possible. It’s not hard to run a 70b model locally, but the queue can get backed up with multiple users unless you have very strong hardware. Still, you can shift overflow to the cloud if you want.
You release the game with a variable in game experience. If the player has two 4090s chugging away, she runs everything locally. If he’s got an rx480, a barebones 1b model or a subscription for the nicer AI NPCs, which can open up AI driven side quests (adding minor content within having to write it). Include a “free” month when you register the game.
Enshitification at its best. This is all part of the inexorable current sweeping us along to the post capitalist society, where corporations no longer sell anything but rather rent or sell access to their things. We’re already there with phones, most consumer computers, farming equipment, cars, and gradually, more and more consumer goods.
With pervasive automation, we are accelerating towards a future where money is meaningless, but not in a nice humanistic star-trek kind of way… more in a dystopian, no need to pay wages to anyone because automation, so we just need land, natural resources, and energy kind of way.
It’s grey goo, on a macro scale so you have to get into space to see it for what it is.
The current pushes us towards a time, soon, when power is the only currency that matters, and justice is reduced to the will of the stronger.
If we want to have something better than trying to compete for resources as squishy humans alongside technofacist enclaves where humans are sparse and wield unprecedented power through massive robotic capabilities, we need to start making changes now.
It’s not going to get better any time soon unfortunately. The technocrats have fully captured the US government. Turns out cyberpunk was the most accurate sci-fi portrayal of our future.
What is there of "post capitalist" in all of that? I see capital and scarcity all the way down. Maybe scarcity is artificially enforced on things that wouldn't be scarce anymore but nothing changed substantially compared to 100 years ago.
What changed is that “the people” are not serfs, they are irrelevant competitors for resources. The only thing you can do with money is pay people for their time. In the end wages is the end consumer of currency. It serves no other purpose.
We are entering a new chapter, where money will cease to be relevant at all. Only land and energy will be relevant. The elite will not need people at all anymore, people and society in general will become annoyances, at best. Everything is better when you have less people to share it with, so “depopulation” will likely be in vogue.
Technofacist enclaves will have the monopoly of coercive force, and will probably fight amongst themselves for resources, but the population of people outside those entities will be a lot like ants.
I thought it meant the change of everything to pay to play… perhaps the meaning is shifting from its original purpose and is expanding to encompass the ruining of everything by corporations in the name of greed? Idk. If seen it used a lot in that context in the last 6 months. I also thought it was just online decay, but I’ve seen it used so much in a broader sense that I think, as a new word, it’s still finding its footing in the greater lexicon.
This is kind of like a self generating agentic context.. cool. I think regular agents, especially adversarial agents, are easier to get focused on most types of problems though.
Off -prem AI is just publishing proprietary or private data with extra steps. If it can’t be done on-prem with a model I can download, I’m not going to hook it into any of my workflows. Period .
Fortunately, on-prem models with agents bolted on are good enough for most things. For the rest I’ll just cut and paste into the web interface on off premises services. That way I know what I’m sharing.
Give it a basic postscript printer driver and a microSD slot for future expansion, a gopher client and a mud client, maybe access to a couple other legacy protocols but no standard web access so it won’t work as a media consumption device or social media outside of things like (AI?) moderated MUDs or chatrooms. You’d want a python or JavaScript interpreter and a simple file manager. Also perhaps an interface to Wikipedia and a telnet type interface to a chatbot API for local or commercial models.
It runs a little against my grain, but perhaps to enable the network Stuff It would need a PPPOE connection to a gateway so that users could be registered and controlled to remove bad actors : tied to chip ID for a “secured area” that parents could opt in to ? iDK, online safety is a tough problem for children. Maybe there is a better solution than total AI surveillance and access controls, but how would you keep bad actors out? Or maybe just not offer any online functionality at all, except maybe a Bluetooth proximity based link to other machines to enable LAN parties?
Something like that would make a pretty great kids computer that could give them access to a complete k-6 education and much deeper computer skills than the current host of consumption oriented devices.
reply