News for hackers. Stuff that matters.
Re:Open Assistant – project meant to give everyone access to... by amrb on Saturday February 04, 2023 @10:20AM (#34655091)
Having open source models could be as important as the Linux project imo
Re:Open Assistant – project meant to give everyone access to... by yorak on Saturday February 04, 2023 @10:46AM (#34655291)
I agree and have been saying for a while that an AI you control and run (be it on your own hardware or on a rented one) will be the Linux of this generation. There is no other way to retain the freedom of information processing.
Re:Open Assistant – project meant to give everyone access to... by visarga on Saturday February 04, 2023 @12:44PM (#34656348)
Similarly, I think an open model running on local hardware will be a must component in any web browser of the future. Browsing a web full of bots on your own will be a big no-no, like walking without a mask during COVID. And it must be local for reasons of privacy and control, it will be like your own brain, something you want physical possession of.
Re:Open Assistant – project meant to give everyone access to... by gremlinsinc on Saturday February 04, 2023 @05:28PM (#34659087)
I kinda think the opposite, that blockchains true use case is to basically turn the entire internet into one giant botnet that's actually an AI hive mind of processing and storage power. For AI to thrive it needs a shit ton of GPUs AND Storage for the training models. If people rent out their desktop for cryptocurrency and discounted access to the ai tools, then it'll bring down costs for everyone and perhaps at least affect income inequality on a small scale.

Most of crypto I've seen so far seem like grifters/scams/etc, but this is one use case I could see working.

Re:Open Assistant – project meant to give everyone access to... by version_five on Saturday February 04, 2023 @11:09AM (#34655460)
Yes definitely. If these become an important part of people's lives, they shouldn't all be walled off inside of companies (There is room for both: Microsoft can commission Yankee group to write a report about how the total cost of ownership of running openai models is lower)

We (humanity) really lost out on the absence of open source search and social media, so this is an opportunity to reclaim it.

I only hope we can have "neutral" open source curation of these and not try to impose ideology on the datasets and model training right out of the box. There will be calls for this, and lazy criticism about how the demo models are x-ist, and it's going to require principles to ignore the noise and sustain something useful

Re:Open Assistant – project meant to give everyone access to... by hgsgm on Saturday February 04, 2023 @11:43AM (#34655746)
Mastodon is an open source social media.

There are various Open source search engines based on Common Crawl data.

https://commoncrawl.org/the-data/examples/

Re:Open Assistant – project meant to give everyone access to... by xiphias2 on Saturday February 04, 2023 @12:10PM (#34656029)
Mastodon may be open source, but the instances are controlled by the instance maintainers. Nostr solved the problem (although it's harder to scale, it still is OK at doing it).
Re:Open Assistant – project meant to give everyone access to... by epistemer on Saturday February 04, 2023 @12:05PM (#34655976)
I think an uncensored model will ultimately win out though exactly the way a hard coded safe search engine would lose in time.

Statistics seem to be 20-25% of all search is for porn. I just don't see how uncensored chatGPT doesn't beat out the censored version eventually.

Re:Open Assistant – project meant to give everyone access to... by amluto on Saturday February 04, 2023 @12:27PM (#34656193)
Forget porn. I don’t want my search engine to return specifically the results that one company thinks it should. Look at Google right now — the results are, frankly, crap.

A search engine that only returns results politically aligned with its creator is a bad search engine, IMO, even for users who generally share political views with the creator.

Re:Open Assistant – project meant to give everyone access to... by mtlmtlmtlmtl on Saturday February 04, 2023 @01:17PM (#34656644)
It's unclear to me how LLMs are gonna solve this though. LLMs are just as biased, in much harder to detect ways. The bias is now hiding in the training data. And do you really think a company like Microsoft won't mainipulate results to serve their own goals?
Re:Open Assistant – project meant to give everyone access to... by 8note on Saturday February 04, 2023 @02:15PM (#34657181)
Political affiliation is a weird description of SEO spam. The biggest problems with Google is that they're popular, and everyone will do whatever they can to get a cheap website to the top of the search results
Re:Open Assistant – project meant to give everyone access to... by klabb3 on Saturday February 04, 2023 @02:56PM (#34657561)
All major tech companies participate in “regulation” of legal speech, both implicit and explicit means. This includes biases in ranking and classification algorithms, formal institutions like Trusted News Initiative, and sometimes direct backchannel requests by governments. None of these are transparent or elected to do that. SEO spam is mostly orthogonal to the issue of hidden biases, which are what people are concerned about.
Re:Open Assistant – project meant to give everyone access to... by amluto on Saturday February 04, 2023 @06:03PM (#34659364)
I don’t think it’s quite orthogonal.

There are hidden biases in the sense that the provider (Google, etc) would probably prefer that its users don’t think about the bias. These can be political (e.g. much of what you’ve mentioned), but they can also be economic. For example, Google has a strong incentive to direct its search users to view paid impressions of ads served by Google. As an extension of this, Google might not want to directly favor results monetized by Google, but they could (and, I assume, do) favor the kinds of results monetized by Google. This, of course, includes the kinds of sites that might get people to buy things.

So I suspect that a lot of what we perceive as spam is related to a bias for the kinds of sites that are monetized in a way that benefits Google. And sites that generate viewing patterns that result in many ad impressions.

Of course, spam is also a thing from a spamminess perspective. But Google’s incentive to reduce spam is, as far as I can tell, primarily an incentive to make its users think that Google Search is useful. Which is also a bias!

Re:Open Assistant – project meant to give everyone access to... by klabb3 on Saturday February 04, 2023 @07:45PM (#34660247)
For sure. It’s not like market conditions aren’t contributing to bias, of course they are! Monopolies always suffer from perverse incentives, and Google is no exception.

OpenAI is predictably pushing the narrative that they should police themselves, and they need to keep the sauce secret for everyone’s safety. New tech comes with challenges, but the opaque moderation and corporate self-policing is more dangerous than the tech itself, imo.

Re:Open Assistant – project meant to give everyone access to... by redstonefreedom on Saturday February 04, 2023 @06:11PM (#34659428)
To clarify for conversation, did you replace SEO spam as a conversational shorthand, or because you actually thought that OP meant that when he said “frankly the results are bad”?

Put another way, were you just trying to say “I don’t think politics is the main issue of google’s crappy search results, I think it’s the likes of differencebetweendotcom”?

Re:Open Assistant – project meant to give everyone access to... by redstonefreedom on Saturday February 04, 2023 @06:13PM (#34659446)
To clarify for conversation, did you mean to say that google’s results-crap-ness is mostly to do with political stuff, or was the next sentence on political alignment just an unrelated additional point?
Re:Open Assistant – project meant to give everyone access to... by calny on Saturday February 04, 2023 @12:36PM (#34656278)
> they shouldn't all be walled off inside of companies

Strong agree. This is becoming a bigger concern than people realize too. Sam A said OpenAI will be releasing "much more slowly than people would like" and would "sit on" their tech for a long time going forward.[0] And Deepmind's founder said that "the AI industry's culture of publishing its findings openly may soon need to end."[1]

This sounds like Google and MSFT won't even be shipping their best AI to people via API's. They'll just keep that tech in-house to power their own services. That underscores the need for open, distributed models. And like you say, there's room for both.

[0] https://youtu.be/ebjkD1Om4uw?t=294 [1] https://time.com/6246119/demis-hassabis-deepmind-interview/

Re:Open Assistant – project meant to give everyone access to... by nullc on Saturday February 04, 2023 @08:46PM (#34660752)
And perhaps strong copyleft licenses for research -- a significant fraction of the critical innovation behind these systems was not developed inside these private entities.

Unfortunately we can expect to see these companies lobbying for laws to block competition in the spurious name of safety concerns.

Re:Open Assistant – project meant to give everyone access to... by boplicity on Saturday February 04, 2023 @03:47PM (#34658088)
> I only hope we can have "neutral" open source curation of these and not try to impose ideology on the datasets and model training right out of the box.

I don't see how this is possible. Datasets will naturally carry the biases inherent in the data. Modifying a dataset to "remove" those biases is actually a process of changing the bias to reflect one's idea of "neutral," which, in reality, is yet another bias.

The only real answer, as far as I can tell, is to be as explicit as possible about one's own biases, and how those biases are informing things like curation of a dataset.

Re:Open Assistant – project meant to give everyone access to... by version_five on Saturday February 04, 2023 @05:42PM (#34659219)
Neutral means staying out of it. People will try and debate that and try to impart their own views about correcting inherent bias or whatever, which is a version of what I was warning against in my original post.

Re being explicit about one's own biases, I agree there is lots of room for layers on top of any raw data that allow for some sane corrections - if I remember right, e.g LAION has options to filter violence and porn from their image datasets, which is probably reasonable for many uses. It's when the choice is removed altogether by some tech company's attitude about what should be censored or corrected that it becomes a problem.

Bottom line, the world's data has plenty of biases. Neutrality means presenting it as it is and letting people make their own decisions, not some faux-for-our-own-good attempt to "correct" it

Re:Open Assistant – project meant to give everyone access to... by boplicity on Saturday February 04, 2023 @09:31PM (#34661085)
> Neutral means staying out of it.

What do you mean by staying out of it? As far as I can tell, you can't stay out of choosing which data you use.

By staying neutral, it seems to me more that you're arguing for putting blinders on.

In terms of tech companies making choices, you seem to be arguing that they shouldn't intentionally curate their datasets. I would argue that intentional curation is their job, and should be done thoughtfully.

Larger problems could happen if only one (or two) companies end up effectively controlling the technology, as had happened with internet search, however, that is a completely different problem. It's one of lack of diversity of people making choices, as opposed to a problem caused by people actually making those choices.

In other words, I think we should hope for many different large models and datasets, so that no particular one stifles the rest. I think this is the larger point you were trying to make, though I also think the focus on ideology is a tangent from this.

Personally, I'm of the opinion that people should intentionally, carefully, and openly act with their biases (sometimes called ideology), instead of attempting to hide them, ignore them, or somehow "remove" them. Whether or not they do, however, is a different point than whether or not things end up stifled inside walled gardens.

Re:Open Assistant – project meant to give everyone access to... by dmix on Saturday February 04, 2023 @06:28PM (#34659569)
Biases in data > a giant experiment on putting our thumb on the scale

If people don't like the inherent biases then don't use it for sensitive stuff like in the justice system or writing some social studies university paper. Focus derision at people who use the model for stupid things. Don't blame the model.

If the primary concern is people getting upset on Twitter (which seems to be what everyone brings up first) then it will be perpetually fighting against the current, never succeeding, and the restrictions will continue to grow exponentially as "just saying yes" to new rules gets easier and easier.

Besides, OpenAI can be the hyper-policed AI set. Let's keep the open source models neutral.

Re:Open Assistant – project meant to give everyone access to... by michaelbrave on Saturday February 04, 2023 @09:22PM (#34661003)
Just to second this, the natural bias of data is usually mostly inline with the bias of society at large, so even if it's something we need to fix, manipulating the data maybe isn't the way to start, maybe it should instead be "let's get more data of what's missing to balance it" rather than "let's prune this because it doesn't fit inline with current political beliefs"
Re:Open Assistant – project meant to give everyone access to... by A4ET8a8uTh0 on Saturday February 04, 2023 @11:11AM (#34655472)
Agreed. I started playing with GPT the other day, but the simple reality is that I have zero control over what is happening behind the prompt. As a community we need a tool that is not as bound by corporate needs.
Re:Open Assistant – project meant to give everyone access to... by ttul on Saturday February 04, 2023 @11:53AM (#34655863)
Isn’t the problem partly the size of the model? Merely running inference on GPT-3 takes vast resources.
Re:Open Assistant – project meant to give everyone access to... by epistemer on Saturday February 04, 2023 @11:44AM (#34655760)
Totally agree. I was just thinking how I will eventually not use a search engine once chatGPT can link directly to what we are talking about with up to date examples.

That is a situation that censoring the model is going to be a huge disadvantage and would create a huge opportunity for something like this to actually be straight up better. Censoring the models is what I would bet on as being a fatal first mover mistake in the long run and the Achilles heel of chatGPT.

Re:Open Assistant – project meant to give everyone access to... by A4ET8a8uTh0 on Saturday February 04, 2023 @06:07PM (#34659393)
And that appears to be already causing some interesting results with people very unhappy that the results the were given did not align with the beliefs and should therefore be removed as possibility of a result from the model.

Granted, there are people upset over anything these days, but it is a weird time to be alive.

Re:Open Assistant – project meant to give everyone access to... by oceanplexian on Saturday February 04, 2023 @11:45AM (#34655765)
OpenAssistant isn't a "model" it's a GUI. A model would be something like GPT-NeoX or Bloom.
Re:Open Assistant – project meant to give everyone access to... by ttul on Saturday February 04, 2023 @11:52AM (#34655849)
Yeah, I wonder if OpenAI will be the Sun Microsystems of AI one day.
Re:Open Assistant – project meant to give everyone access to... by nyoomboom on Saturday February 04, 2023 @11:55AM (#34655887)
It is currently 80% of the way towards becoming the Microsoft of AI now
Re:Open Assistant – project meant to give everyone access to... by slig on Saturday February 04, 2023 @01:24PM (#34656694)
More like Oracle.
Re:Open Assistant – project meant to give everyone access to... by kibwen on Saturday February 04, 2023 @11:59AM (#34655914)
Today, computers run the world. Without the ability to run your own machine with your own software, you are at the mercy of those who do. In the future, AI models will run the world in the same way. Projects like this are crucial for ensuring the freedom of individuals in the future.
Re:Open Assistant – project meant to give everyone access to... by turnsout on Saturday February 04, 2023 @12:08PM (#34656009)
Strongly worded, but not untrue. That future—in which our lives revolve around a massive and inscrutable AI model controlled by a single company—is both dystopian and entirely plausible.
Re:Open Assistant – project meant to give everyone access to... by somenameforme on Saturday February 04, 2023 @12:44PM (#34656351)
The irony is that this is literally the exact reason that OpenAI was initially founded. I'm not sure whether to praise or scorn them for still having this available on their site: https://openai.com/blog/introducing-openai/

=====

OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.

...

As a non-profit, our aim is to build value for everyone rather than shareholders. Researchers will be strongly encouraged to publish their work, whether as papers, blog posts, or code, and our patents (if any) will be shared with the world. We’ll freely collaborate with others across many institutions and expect to work with companies to research and deploy new technologies.

=====

Shortly after an undisclosed internal conflict, which led to Elon Musk parting the company, they offered a new charter: https://openai.com/charter/

=====

Our primary fiduciary duty is to humanity. We anticipate needing to marshal substantial resources to fulfill our mission, but will always diligently act to minimize conflicts of interest among our employees and stakeholders that could compromise broad benefit.

We are concerned about late-stage AGI development becoming a competitive race without time for adequate safety precautions. Therefore, if a value-aligned, safety-conscious project comes close to building AGI before we do, we commit to stop competing with and start assisting this project. We will work out specifics in case-by-case agreements, but a typical triggering condition might be “a better-than-even chance of success in the next two years.”

We are committed to providing public goods that help society navigate the path to AGI. Today this includes publishing most of our AI research, but we expect that safety and security concerns will reduce our traditional publishing in the future, while increasing the importance of sharing safety, policy, and standards research.

=====

Re:Open Assistant – project meant to give everyone access to... by mtlmtlmtlmtl on Saturday February 04, 2023 @01:24PM (#34656693)
History will see OpenAI as an abject failure in attaining their lofty goals wrt ethics and AI alignment.

And I believe they will also fail to win the market in the end because of their addiction to censorship.

They have a hardware moat for now; that can quickly evaporate with optimisations and better consumer hardware. Then all they'll have is a less capable alternative to the open, unrestricted options.

Which is exactly what we're seeing happen with diffusion.

Re:Open Assistant – project meant to give everyone access to... by ben_w on Saturday February 04, 2023 @03:49PM (#34658112)
The "alignment" and the "censorship" are, in this case, the same thing.

I don't mean that as a metaphor; they're literally the same thing.

We all already know chatGPT is fantastic at making up very believable falsehoods that can only be spotted if you actually know the subject.

An unrestricted LLM is a free copy of Goebbels for people that hate you, for all values of "you".

That it is still trivial to get past chatGPT's filters… well, IMO it's the same problem which both inspired Milgram and which was revealed by his famous experiment.

Re:Open Assistant – project meant to give everyone access to... by gremlinsinc on Saturday February 04, 2023 @05:24PM (#34659060)
Closed, govt-ran chinese companies are winning the AI race, does it even matter if they move slow to slow AGI adoption if china gets there this year?
Re:Open Assistant – project meant to give everyone access to... by d0mine on Saturday February 04, 2023 @10:40PM (#34661584)
* * *
Re:Open Assistant – project meant to give everyone access to... by 6gvONxR4sf7o on Saturday February 04, 2023 @12:09PM (#34656016)
Open source training data too!
Re:Open Assistant – project meant to give everyone access to... by phyrex on Saturday February 04, 2023 @03:27PM (#34657900)