Oh boy, how surprising.
The bait and switch classic.
I’m clutching my pearls as I type this.
So the development of inorganic intelligence, considered by many as an inflection point in human civilisation is to be handed to business graduates who are historically proven to be capable of any level of atrocity in the name of corporate greed. America, fuck yeah.
AmericaGreed, fuck yeah.Don’t fool yourself. The USA lost the exclusivity deal on unchecked corpo greed a long time ago. This is a global issue now.
Always has been.
Yeah, the American tag was just a throwaway line, greed unchecked, insane and self-harming has always been with us. We let it sit with us around our camp fires like wolves but unlike wolves we never tamed it.
Then again, the US and China are basically the only players in this “game” atm. Hugging Face is trying hard to get the EU on-boarded, and I’m sure we’ll see more contenders. But right now it’s a 2-player game.
I think western interests at least are beginning to detach from nation states.
Actually corporations themselves are 99% of what people fear about AGI already in their inhuman decisionmaking to the detriment of humanity.
What do you mean by “inorganic intelligence,” exactly? Do you think openai has already achieved it?
As if any other group did not prove that just as much, if not more so.
Removed by mod
“ClosedAI” rebrand when?
🤣
NopeAI
Open Your Wallet AI
Stop depending on these proprietary LLMs. Go to !localllama@sh.itjust.works.
There are open-source LLMs you can run on your own computer if you have a powerful GPU. Models like OLMo and Falcon are made by true non-profits and universities, and they reach GPT-3.5 level of capability.
There are also open-weight models that you can run locally and fine-tune to your liking (although these don’t have open-source training data or code). The best of these (Alibaba’s Qwen, Meta’s llama, Mistral, Deepseek, etc.) match and sometimes exceed GPT 4o capabilities.
And there are also free, online hosted instances of those same LLMs in a (relatively speaking) privacy-protecting format from DuckDuckGo, for anyone who doesn’t have a powerful GPU :)
Interesting. So they mix the requests between all DDG users before sending them to “underlying model providers”. The providers like OAI and Anthropic will likely log the requests, but mixing is still a big step forward. My question is what do they do with the open-weight models? Do they also use some external inference provider that may log the requests? Or does DDG control the inference process?
All requests are proxied through DuckDuckGo, and all personalized user metadata is removed. (e.g. IPs, any sort of user/session ID, etc)
They have direct agreements to not train on or store user data, (the training part is specifically relevant to OpenAI & Anthropic) with a requirement they delete all information once no longer necessary (specifically for providing responses) within 30 days.
For the Llama & Mixtral models, they host them on together.ai (an LLM-focused cloud platform) but that has the same data privacy requirements as OpenAI and Anthropic.
Recent chats that are saved for later are stored locally (instead of on their servers) and after 30 conversations, the last chat before that is automatically purged from your device.
Obviously there’s less technical privacy guarantees than a local model, but for when it’s not practical or possible, I’ve found it’s a good option.
Okay that sounds like the best one could get without self-hosting. Shame they don’t have the latest open-weight models, but I’ll try it out nonetheless.
i’m not so sure on the privacy of any of this.
The issue with that method, as you’ve noted, is that it prevents people with less powerful computers from running local LLMs. There are a few models that would be able to run on an underpowered machine, such as TinyLlama; but most users want a model that can do a plethora of tasks efficiently like ChatGPT can, I daresay. For people who have such hardware limitations, I believe the only option is relying on models that can be accessed online.
For that, I would recommend Mistral’s Mixtral models (https://chat.mistral.ai/) and the surfeit of models available on Poe AI’s platform (https://poe.com/). Particularly, I use Poe for interacting with the surprising diversity of Llama models they have available on the website.
There are open-source LLMs you can run on your own computer if you have a powerful GPU.
What defines powerful? What if you don’t have the necessary hardware?
You can check Hugging Face’s website for specific requirements. I will warn you that lot of home machines don’t fit the minimum requirements for a lot of models available there. There is TinyLlama and it can run on most underpowered machines, but its functionalities are very limited and it would lack a lot as an everyday AI Chatbot. You can check my other comment too for other options.
llama is good and I’m looking forward to trying deepseek 3, but the big issue is that those are the frontier open source models while 4o is no longer openai’s best performing model, they just dropped o3 (god they are literally as bad as microsoft at naming) which shows in benchmarks tremendous progress in reasoning
When running llama locally I appreciate the matched capabilities like structured output, but it is objectively significantly worse than openai’s models. I would like to support open source models and use them exclusively but dang it’s hard to give up the results
I suppose one way to start for me would be dropping cursor and copilot in favor of their open source equivalents, but switching my business to use llama is a hard pill to swallow
I thought they were a for-profit company all this time.
No problem, after they release all the data collected under the excuse of public good and progress.
Sam:
That’s very open of them
Open to All Income.
deleted by creator
Weird, I said this shit for years, and I was upvoted into the heavens, agreed with, called a hero, and acknowledged as a result.
Maybe is not what was being said?
deleted by creator
Heh, I warn about Mozilla/Firefox all the time and get the same. I hope I’m wrong though :(
Everything was clear about Mozilla the moment they started fighting the ecosystem around Gecko, with alternative browsers, useful extensions and so on. And, of course, the old usable UI.
People just forget what they don’t know how to process.
Disagree, XUL was a dead end that either needed shooting behind the bike shed or it’d have taken Mozilla down with it inevitably. It froze their internal architecture to a design that didn’t care about multicore or modern security. Switching to a proper extension api (it didn’t matter if it was chromes or their own, only that they are willing to make their own decisions, like in manifest v3).
That said, I suspect the real death blow was when they killed servo, that project was their distant salvation, a chance to genuinely outcompete technologically and direct where browsers need to go next. I too hope I’m wrong and they can figure out a path forward, but they’ve shown little ambition from the top, so I’m not holding my breath.
Edit: you could argue that the solution to XUL should have been an upgrade to modern design rather than death, but that would have just been an expensive temporary reprieve, the world doesn’t stop changing, it was always going to be slow to correct to whatever direction they needed to go next (and meanwhile every extension dev would be screaming murder every time they killed some braindead api designed 20 years ago).
XUL itself - of course.
Edit: you could argue that the solution to XUL should have been an upgrade to modern design rather than death, but that would have just been an expensive temporary reprieve, the world doesn’t stop changing,
I’m not sure what do you mean by that. No deep customization at all is, of course, easier to support than some.
I don’t care about preserving the feel of XUL, or any aesthetics, but I do care about its role.
It’s not about specific extensions and specific language. It’s about the “before” allowing things like Conkeror and any kind of appearance change conceivable and the “after” - not, if we don’t count stupid CSS that breaks with every update.
Booooooooooo!
Anyway: ill just keep using alpaca to run llms locally
is there an easy way to do this that doesn’t require me to understand how github works?
I think that in that case, YouTube is your friend. There are a few pretty straight forward videos that can help you out; if you’re serious about it you’re going have to, eventually, become familiar with it.
I recommend Ollama, its easy to setup and the cli can download and run llms. With some more techsavviness you can get openwebui as a nice ui.
For someone who doesn’t understand GitHub, the CLI might be a bit much, FWIW.
It would be nice if there were a GUI, download-and-run single click app with a webui built in.
Alpaca for linux is easy to use. You just install the flatpak and the llm of your choice. You dont need to know how to use github. (It might have a windows version but im not sure)
So, ads in chat now?
‘subtle’ product recommendations
Yup, conversational product plugs
They’ve already started testing that at Google For ad enhancement and For immersive ads there’s no way they keep the chatting models pristine and ad-free
The dystopian future of “pay to use this miraculous product or it will shove advertisements down your throat in a way we know will work because we’ve trained it to sell specifically to you”
They should also change their name to ClosedAI while they’re at it.
Hahaha. April 1st is early this year.
They are never going to make enough money by selling licenses and subscriptions for the cost of their current models (smarter people than me have made good estimates), let alone the future ones. Those future models are at a much worse performance-cost ratio. Ads will at best bring in about 1 usd per user per month (estimated by Facebook revenue and number of users) - double or triple it just for lolz, and they would still be losing money.
So… how will this be pulled off? Only wrong answers!Have a partnership with Microsoft and ship Windows 12 as the new “AI only” OS. Every command must go through ChatGPT to work. Then push updates to older Win11 OS to make them unusable.
From what I’ve heard, they don’t need to push updates to achieve that
How fast are they burning money right now?
Based on their funding rounds, $10 billion lasts about 18 months.
So about $555 million per month.
They don’t care if they earn money the next 5-7 years.
And they will hit the point of a great model doing human work for less than a monthly salary. It’s just a matter of time.
Without enough funding, they absolutely will care.
Thats between $33 billion and $47 billion at current costs. Someone needs to fund that.
I’d also note that their models seem to be getting worse, with outright irrelevant answers, worse perfoemance, failures in following instructions, etc. Stanford and UC Berkeley did a months-long comparison, and even basic math is going downhill.
I’d rather say that it’s a matter of exponentially increasing funding and computing power.
I am honestly very very curious: how?
They’ll upgrade the Aibo and stick Altman’s face on it. People in offices can enjoy kicking it.
There was never another outcome.
Capitalism breeds one thing, and it certainly isn’t innovation, and it most definitely isnt not-for-profit innovation.
Capitalism is extremely good at breeding superficial, go-to-market innovation. It’s less good at funding the pure research that leads to major discoveries. But once it gets closer to engineering than to science, it’s highly effective. Even Marx commented on that.
No kidding. 🙀
