I run local models. They're fun to play with. I get a bit of a dopamine hit when it works.
They're selling addiction. This is fucking disgusting.
I can’t.
If a government thinks that ChatGPT even in its current form is a big boon, it makes sense to do this.
This is also insurance - your population gains LLM literacy and will have an advantage over other countries; even if it's only a tiny bit, at the level of states it adds up.
No, not gonna pay yet another tax, mr. Taxman.
My analogy is using AI is like using a navigation system, you can end up delegating everything to it and drive into a river...
So it’s basically I giant government-sponsored free trial.
snort
That is how AI boosterism works here.
Tech employees worried their stock will drop in value (laugh emoji)
Next, force an eyeball scan on the peasant population.
On the taxpayer money.
If you force me to pay taxes, and you offer me free access to inference, I don't see why I would run my local, privacy-focus model.
Malta has a population of only 550k.
Everyone in Malta could already, before this deal/plan, and even without it now, use ChatGPT (or any other LLM model/service, whether free or premium.)
I'm Maltese so feel free to be as detailed as needed.
What’s dangerous here is people will eventually stop thinking critically altogether, about anything, and their view of the world will be based on what they are shown on AI apps.
But that’s besides the point, the whole initiative is self-defeating by design. This isn’t like power, it’s something humans do inherently possess, this is simply a way to amplify what already exists. Intelligent people using AI generally seem to be more productive than when they don’t use it, and lazy or unintelligent people generally see cognitive decline, at least based on what I’ve heard online but I could be wrong on that.
So saying “this is where you get intelligence” is both false marketing and destructive to OpenAI as a company, since by all definitions, it isn’t true.
For OpenAI because they get a lot of money and and for the government because they can keep tabs on how people use LLMs to make sure they're not doing anything naughty.
EU created software like Mistral is no different
come on now, surely you're not serious.
Your body also generates electricity and natural gas. Do you also get upset when energy companies claim to provide these services as a utility?
Does AI actually provide intelligence?
We do and we don't. If you would go out there and talk to a random person about elliptic curves and matrix multiplications and whether you hit a performance ceiling in a specific 2x2 multiplication thingy with Karatsuba and wnaf, they would not know half the words, but the lying and flattering machine will be able to hold the conversation.
The thing will not get all things right and bullshit me about DSTU4145 using normal basis, will lie about A being set to 1 for all standard curves, but it's definitely more intelligence that you can get from a taxi driver.
If it's not general superintelligence right there for five bucks a piece, I don't know what is
Is a calculator intelligent? I can 'talk' to it via pushing buttons.
Wikipedia has existed for decades...
Lol, they are literally just promising to make people fungible. Tale as old as time.
More like off the cliff!
I mean, what's the point of this question even. The thing is either useful or fun or it's not. I personally think the whole AI is the work of devil tempting us, but some people would say that about pork sausages and Paulaner and I like my pork sausages with Paulaner.
So the fact that you get it free after doing some basic due diligence is actually a big deal in the local context.
I suspect in a few years it's going to be strange to talk to him and other people there. It's already hard to explain to people that "Yeah, you can have a phone call and it can sound like your dad but it might just be a chat bot."
If I never have to hear anything about AI ever again it will be too soon
Nobody is obligated to use it. It just moves the price to $0 for people in Malta who choose to use it. Same service.
Of course Maltese citizens can still choose not to use ChatGPT (until it becomes mandatory), but if the State supported education is bound to one particular tool, storing user data outside the EU’s jurisdiction, I think that’s something to discuss.
> How do these EU citizens submit a request for all their personal data to be deleted from OpenAI records
Probably by sending an e-mail to a designated address, like most services that operate in the EU, but you can read their TOS if you'd like to be sure.
Care to elaborate or we have become completely apathetic to any display of sleaze?
If you want my commentary on the political context, obviously I think it's not very intelligent for nations to be trusting a US corporation with all of their citizens' data. I think the most impactful use of LLMs is going to be their usage as surveillance and propaganda tools, so this is probably not a prudent decision. But legally, as pertains to GDPR, this is not different from the status quo in any way.
so: I doubt anyone has to care about that pesky GDPR if they buy the government of Malta.
I got a spot. We were shown how to copy and paste data from excel and other data sources into the chat interface. We had sample data to work with, there was always someone in class who would say "mine didn't work." The developers in the room asked about codex, the instructor said she wasn't a developer.
We did get a certificate though. There was nothing they could teach that you couldn't learn by using the free version in your own time. Whatever they are doing with the Maltese government is just to increase the monthly active user count.
But the people in charge just want the employees to just answer some questions so they can handover Claude or Chat GPT licenses so they can show people are using AI to improve productivity.
There are people who don’t know when to use AI and when not to use Ai and think they can just Claude their way through everything. I wanted to change that but when the whole idea is to just increase AI use I guess they don’t care about how AI is used.
I will not address the things he pitched (as coming soon), as I'm a developer and (hopefully) not the target audience, but I was quiet surprised when they made a questioneer asking how many people use ai and how frequently. (The target demographic was middle management, product owners etc)
75% of people answering said they're using it daily and considered it an essential tool they need to work
Considering it was anonymous I was expecting lower numbers, honestly.
Why not?
The thing I have a real issue with, and which seems more common, is the belief that they can cut raises because AI will make them more productive. In that case, the best employees (read: those most capable of leveraging AI effectively) will leave to find better paying work and the remainder will be too busy with the additional workload to have time to figure out how to use AI to make themselves more efficient.
Fuck all of this.
I jumped at the chance to not have to be in the OR from 7am-5pm doing the same old same old but instead relax and learn something useful.
Bad choice.
The instructor and material were deadeningly boring; I couldn't even begin to enter into the computer the right search request format and terms and as I sat there I was reminded of my days in elementary school watching the big round clock on the wall tick away the minutes until the final dismissal bell.
Because our chairman was in the class and had encouraged all of us tenure-track faculty to take the course, I couldn't bail after the first day but had to return for the second day.
Subsequently I continued using the biomedical librarian to request my searches (it took just a couple minutes to fill out the form) with excellent results.
https://www.npr.org/2012/05/03/151860154/put-away-the-bell-c...
I asked if he had tried out Claude code or anything similar. His answer: My company has scheduled a training course in that so I'll wait
:(
That's the hint. Most companies >50 employees suck.
As someone who never bothered to get any certificates (beyond a University degree) even when I'd do online courses (of which the most course-like must've been fast.ai), are these ever actually useful in any manner?
Many of them you can simply take the exam over and over until you pass, and then stick a shiny stupid badge on your LinkedIn profile.
In our case, we get our entire team AWS solution architect certs as well just so we can always tell our customers that our whole team is certified (we do a lot of “forward deployed” stuff for enterprise customers).
Grnnnnnnnnnnnnnnnnnnguuurnnngh.
I remember the copy and paste drudgery from the early days of ChatGPT. It was a miserable and joyless experience. Nowadays (and for a long time) you can simply attach the file.
It does take time and a little skill to know the edges of the AI tools. What's reasonable? What's not? What's likely to hallucinate? You could get something in the rough bounds of trust.
I can see a class helping with that.
These violations come with hefty fines.
Companies and employees always make their decisions based on a risk/reward basis.
Sometimes a commercial contract (like Microsoft Copilot) is enough to cover your ass and to meet the needs of the regulator.
Even if the operator is exactly the same.
Laws are constraints to navigate, but if you are successful enough (ahem, rich) then they don’t apply to you.
At the moment what the EU wants is to make sure that in the long-term they can access your private information.
Realistically if you are in the EU you have more risks telling your darkest secrets to a EU-hosted model that the government will arrest you, than to a Chinese-model (who doesn’t collaborate).
EU Chat Control, is here to protect kids and protect you from terrorists; you don’t want to claim you support pedophiles right ?
So following these rules is always a matter of choice.
Respect and you will be stuck with your shitty Mistral and no privacy, not respect and you have your shiny Claude that you have to think what to input inside.
But nobody uses LLMs that aren't Gemini or Copilot enterprise, as they are already on Google cloud or Microsoft offering already.
And there's high pressure on workers to find use cases where AI can boost productivity, with bonuses dangling on who finds real case scenarios.
I don't know about the results of these experiments, but I know unhappiness is widespread.
"Make the AI do xyz"
That clearly needs a custom harness to integrate with ORG tooling.
"No we won't pay for token usage, make it work with the subscription were already paying for"...
Guess you don't want AI then...
Cue up the next cynical bad take.
That's true for almost everything in life.
> Whatever they are doing with the Maltese government is just to increase the monthly active user count.
That's one of their main goals. Another main goal is to also make money. There are a few other main goals.
What do you mean by "just to increase"? Did they try to hide their goal? Was it a secret agenda nobody knows about?
These are some strange tautological comments.
"Malta’s corruption is not just in the heart of government, it’s the entire body"
https://www.theguardian.com/commentisfree/2019/dec/03/malta-...
Could you give examples of a few of these to illustrate the kind of thing we're talking about? It currently feels a bit hand wavy from both sides.
That’s quite a difference to most other European countries, although not all.
https://www.anthropic.com/news/anthropic-and-iceland-announc...
I'm just wondering when the US will finally put boots on the ground in Iran...
Regions can sign deals too…?
EU software also collects your data
At least American software is good
If you’re gonna give up your data, at least use the better American software
Crypto, eSports, Gambling, VPNs, etc, all have deep roots in Malta because its easy, less taxed, less restricted and for the most part, people view Malta as a good place.
Malta is among Europe's leading adopters. The country ranks first for workplace AI usage and third overall for general AI adoption. Beyond AI, Malta also has one of the highest rates of social media usage in Europe.
This initiative is less about AI literacy or OpenAI and more reflective of the Malta government's policy of making technology accessible across all society.
It's not the first time either, in the early 2000s a similar partnership with Microsoft which provided heavily subsidized microsoft office licenses.
As a local the Daphne case has been a sore point and something which makes me very angry and sad. However it's not the only country with debatable press freedom. It's somewhat difficult to know that my country has been reduced to a single bad episode.
Regarding DCG's case: trial of the prime suspect due in July I believe. DCG's tragic case was a one-off. Attacks on democracy are much more common and frequent in other Western countries.
Not sure why a bribe would’ve involved: it’s a win-win situation, especially near elections.
That’s whats happens in two sided markets. Everyone’s the product.
The original adage of “if you’re not paying, you’re the product” doesn’t necessarily rule out the converse. The fact that the grandfather comment made a freudian slip makes it funnier.
EU-created domestic software is no better
Actually it’s worse. You still surrender your data but get a worse user experience
That answers my wondering of why on earth all citizens require a paid service of chatGPT.
If all citizens use ChatGPT as search engine instead of Google, their ability to access information and defend against fake news (as of right now) will be significantly improved and we only shift privacy concerns from Google to Open AI, which does seem like the lesser evil.
> Who can register: > > Maltese citizens and residents > Must have an active eID account > No previous AI knowledge is required
From https://mdia.gov.mt/services/ai-for-all-ai-ghal-kulhadd/
Yes, ChatGPT has a large user base, but Malta's not a particularly meaningful datapoint in terms of the population size.
It's so simple, everyone on HN knows this, but I want to scream at every non-technical person "these tools are not reliable! You can't trust them!"
They are so useful but the failure modes need to be hammered into everyone at the very start of each onboarding.
It would be totally rational as a layperson to just ignore AI and just check in every few months to ask "is it reliable yet".
Malta has a population of 500k. Let's assume 100k people use MaltaGPT daily, and they send an average of 10 messages per day, so roughly 1M messages per day. That averages 694 per minute, but at peak could be 3-5x that, so let's say 3000 per minute. Usage will of course vary by day of week and time of day (they could partner with a Pacific island and share inference hardware).
Those 3000 messages per minute translate to 50 messages per second. Let's say average prompt input is 5k tokens, and output is 500. So 250k tokens per second for prompt processing (let's ignore caching for simplicity) and 25k tokens per second for output decode.
If we take a 500B dense model, that concerts to roughly 1 trillion flops per token. So we need 250 petaflops per second of prompt processing and 25 petaflops for output decode. So 275 PFLOPS in compute.
That may sound like a lot, however a NVIDIA DGX B200 machine (8xB200) has a compute of 144 PFLOPS at FP4. That is assuming 100% efficiency which isn't really possible, and we also need to factor in memory usage which we would be limited more by than compute. So let's say we'd need 10 of them. For an entire country to have a sovereign version of ChatGPT.
The cloud cost to rent one machine is around $50/hour, so that would mean our cluster comes to $4.8m per year. However the list price of a machine is around €400k, so the price to buy the cluster outright would be around €5m (you need the rest of the data center too), with operating costs of around €500k per year.
So per citizen: €10 upfront and €1 per year.
Closed software has no place in government whatsoever. It should all be open and/or locally run, and ideally GPL/AGPL licensed.
It is another level of stupidity/moral failing to use closed software supplied from outside your nation, though.
At least the usage data would stay sovereign even if the model was trained somewhere else.
Thank god people protested against it and made them drop the plan.
Just look at this list of services included in Google's AI Pro subscription[1]. Google took everything it could think any consumer might need and bundled for $20/mo. There's even $10 GCP credit (that you can use for AI API calls).
[1] https://support.google.com/googleone/answer/14534406?hl=en
They import food and water. Malta is very hot during the summer. There's AC unit everywhere and it's a default cooling unit as well, as there's no "European winter" there. Everyone collects rain water and stores it on the roof.
They are one tsunami away from being decimated.
There's one company renting servers and it's full of online casinos, just so the companies meet the regulatory requirement.
Malta is the worst place on earth to have a data center I can think of.
Presumably what you refer to as "money laundering" is the impression that Malta attracts foreign investments by offering regulation for poorly regulated industries and tax incentives. Which is essential to maintain competitiveness as a small state. You'd be surprised to hear that most money laundering in Malta is not tied to the igaming industry at all.
Malta is not a good place for a data center because real-estate is expensive and cooling is expensive.
Many jurisdictions literally force them to put education on the boxes.
All positive comments here come from the financially invested or the near-retirement people who need cognitive assistance and are willing to sell out future generations.
Usually this would require the respective customer to agree to sharing that data with a third party.
I'd be surprised if there weren't already phishing attacks that work by pretending to be a LLM.
I'd think that the country's regressive anti-abortion laws are a bigger stain on its reputation. You can root out corruption. Moving the nation's Overton window towards a less illiberal stance tends to take a few generations.
Original point stands that AI is useful and better than manual searches.
Do you think the average people rigorously query multiple angles and carefully read every one of the google results, to synthesize a well rounded viewpoint on any given topic? No, but with LLMs they can do that in one prompt.
EDIT: Needless to say I loath it and I don't know why.
For the not liking it part, I guess that if someone writes a long text, there are more chances to find at least a point of disagreement than a very short sentence
I believe the logical term "converse" means swapping the conclusion and the condition in a logical statement, ie converse(if A then B) = if B then A
So here the converse would be "if you're the product, you're not paying". Which doesn't exactly make sense to me as a claim to make here. Did you just mean to reinforce your first sentence? In which case, I think you mean "the inverse", not the converse. However, I have only used the word converse in a "formal logic" scope (proofs) so I'm not sure if it has a more flexible meaning in informal language use.
Here is how platforms die: first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves. Then, they die.
This is at the “good to their users” phase - giving free access to an otherwise non-free product is being good to them, right?
I have no numbers to back this up.
[1] https://timesofmalta.com/article/the-15-billion-question-is-...
It’s much more of an issue with devs
That’s quite a difference to most other European countries
The difference being that you personally are not prejudiced against them for no other reason than your ignorance and arrogance.
Your username and comment history suggests it might not be wise to take your word as objective truth.
UK (United Kingdom of Great Britain and Northern Ireland) is a country consisting of several countries and other territories.
In the recent past, my department received an email from on high with a list of people who were yet to complete the "anonymous" survey.
I always assume my work-survey answers are traceable back to me, whether it's via self-doxxing with my answers, tracing links of the rootkit-level MDM software that can record my screen, but they pinky-promise to only use for remote assistance, in case I open a ticket with IT.
Trusting that process to be done well is probably not the greatest plan.
That’s not anonymous at that point. That’s an agenda.
Even when you use a tool like Microsoft Forms, where MS really can’t be bothered to deanonimize users unless 3 letter agencies get involved, it’s still possible to do timestamp matching between the proxy/VPN logs and the submission time.
Asume real anonymity only if the URL is the same for everyone and you can fill the survey from any computer on the internet.
But the explanation for why people overhype AI usage is probably simpler. They want to keep their license because it’s a nice perk. They’ll use it to get the gist of a long email thread without bothering the read the details, to get some meeting minutes without validating if that was actually what was said, to generate some crappy modern equivalent of wordart graphics for their presentations, and feel like the time saved to generate what most time is slop was worth it.
When I worked on this (outside of coding) it was a pain to find a use case that really benefited. These were all niche uses that fit an LLM like a glove. These rest was slop, I could see the usage reports, and the BS self reporting surveys. Everyone inflated the numbers and usage to justify keeping their license.
The job market right now sucks so everyone is really just trying to not be the next cut.
Could still be faked ofc, but I don't think they did.
> (The target demographic was middle management, product owners etc)
This leaves a fairly wide set of options for what "essential" entails.
Do 75% of middle management and product owners actually need AI for their job? Seems unlikely.
Do 75% of middle management and product owners use AI to slop up emails, meeting "summaries", and reports? That's quite possible. Would they declare it to be an "essential tool"? One imagines they are not too fond of actually doing meaningful work.
It's quite easy to get high percentages like this when the AI is involved in make-work and the costs are low if not zero. The moment inference costs go up, most of this usage will evaporate.
But maybe the simplest answer is that most people do use the tools daily now and consider them essential...?
As much as HN would hate to think that
It all rest on the shoulders of responsible manager(s) on how moral they are. Many are not.
Of course if you record created/updated timestamps on both, insert both records in the same order, accidently record the user code in the response data, take backups in between responses, have identifying questions or just don't have that many people responding it's easy/not hard to reverse engineer.
But it's quite possible to do right, I did it quite effectively almost by mistake years ago. Sent a customer survey out with generated codes as identifiers recorded with answers. Before sending reminder emails a script grabbed the codes, marked the customer as responded and wiped the code (so I could just get future responses where code was not null to mark next people off). Although I had timestamps the script meant customers were updated in blocks, there really wasn't any data to link them.
I know because the Boss was not happy he couldn't find out which customer had said what, and I had to point out all the communication (with customers and me) called it an anonymous survey, so why would I have saved them?
So it is possible, just not easy even if you intend it, and it's often not intentional...
I don't trust anonymous surveys either now...
If the participant has to trust the survey creator, then it is not anonymous. The survey creator can link the data.
If the survey creator has to trust the participant, the survey is anonymous. The participant can lie in the survey, lie about participating, or submit the survey multiple times.
Your example was not anonymous. But you did not break the participant's trust, thank you! (Or maybe you are lying.)
Anonymous example: Sending a clean link to people to take the survey. If not enough answers have been received, a reminder can be sent to all, with a clause, that says: "if you have already done it, you can ignore the reminder."