You can already run a large LLM (like sonnet 3.5) locally on CPU with 128GB of ram which is <300 USD, but can be offset by swap space. Obviously, response speed is going to be slower, but I can't imagine people will pay much more than 20 USD for waiting 30-60 seconds longer for a response.
And obviously consumer hardware is already being more optimized for running models locally.
But most of the time, we as engineers don't pick the winners. Some C-Suite executive or middle manager, who isn't very technically inclined, picks the winners, and we as engineers are forced to make it work.
As I don't think a engineer has ever had the chance to choose a company's CRM, the CRM with better marketing would always win over the CRM with better engineering.
Question I would pose is, why should engineer have the decision on a new CRM?
They can provide input regarding e.g. maintainability, but majority of input would come from other stakeholders - users and business unit owning the customers whose relationship we want managed, ideally primarily. And it is somebodys job to take these inputs into collective whole.
It was a mind blowing exercise to me 15 years ago when I was telling my boss how horrible our current installation of some ERP software was, and be asked me what's the user perspective. They log in every day, run financial reports they need, and log out. The system was great from their perspective! They had even less concern for my perspective of poor architecture and suboptimal implementation, than I (at that point) had of their perspective and goals. Thank krishna I didn't make the decisions on the CRM :-)
>why should engineer have the decision on a new CRM
well there's the craftsman argument and then there's the broken windows argument.
the craftsman one if obvious: if you're in a devops/IT role and your job is to manage salesforce, then you should have some input in it as it'll affect you efficiency (aka the profitability of your company). A salesman shouldn't be buying tools for the carpenter without the carpenter's input.
the windoww argument is a bit more superficial but still a factor to consider. I may not be working on mainaining saleforce, but I will need to interface with it for logstics purposes. if it's so inefficient that it becomes a chore to track hours or update documentation or etc. it's going to leave a bunch of broken windows. You can still operate with a broken window, but that part of the building will be a place to avoid. You may even try to work around the CRM wherever possible. Which seems to lose the point of a CRM
Upthread, bayindirh posted half a dozen examples of financially-motivated decisions that were actively, deliberately hostile, sometimes fatal, to the customer. We're not just talking about good-enough fiddly details here.
There are two things wrong with this argument. One, it implies that there isn't a better way to deal with an such a malicious organization. And two, it doesn't acknowledge how such a ban creates an obvious opening for abuse.
Holding up the classic Western ideals of Democracy and Freedom is hard because it is much deeper than simply giving people the freedom to access all information so they can form their own opinions. It also requires that these people are educated and trained to be competent critical thinkers and be able to intelligently form their own opinions. It holds its citizens and the government to high standards and will collapse if these standards aren't met. Accordingly, better education and trust in citizens is the better solution, not banning.
As for the obvious opening for abuse, it doesn't have to be said that every system will eventually be maximally exploited, and creating this opening for exploitation will eventually be exploited as well. It is just a matter of time...
Another poor but useful analogy is fast food. Banning bad media is like consuming fast food. It is quick and easy, and "satisfies" the goal within some basic parameters, but it really does more harm than good in the long term.
The argument in no way implies that there is no other way to deal with malice, only that it is an option. The argument further implies that the decision to censor should not be taken lightly. When censorship is being considered in a democratic society, the decision to do so must be argued and debated. Note that it is perfectly possible to be well-educated and still be taken in by bullshit and false information - it happens all the time. Education is a good, not a nostrum, and durable opponents of truth are also motivated, sophisticated and smart. Democracy is about the means you use to undertake drastic decisions, and in no way rules out the restriction of unseemly behaviour.
Maybe explain why my analogy is not useful - I’ll do the same for you. What I initially said did not in any way imply banning should be quick or easy, while “fast food” is not, in itself, bad for you, but a restricted and monotonous diet of anything can well be.
Content is being banned because the governments don't think the citizens collectively can handle that content. This is the government thinking for the people. I wouldn't have an issue if it was really argued and debated, but I don't really see that happening. It is mostly just the government unilaterally deciding that said content is bad and banning it with no really possibility for debate. It isn't too hard to imagine a situation where content would help citizens but harm the government, and who is really deciding what to ban in such a case.
Countries are seeing how much banning content helps governments in places like China, and want that easy "solution" to their problems instead of doing the difficult thing of properly educating citizens and cultivating free and open discourse.
I guess my analogy is bad because it is debatable if fast food, in itself, is bad for you. My issue with your analogy is that the use of violence is abused by governments so much more than it is every used correctly (if that is even possible) and that all citizens would benefit from governments condemning violence. The reason governments don't condemn violence isn't because their use of violence helps citizens, it is because their use of violence helps themselves. And the same thing goes for banning content. So I guess your analogy is actually pretty good,
just counterproductive to your point.
Your analogy was bad in that it represented a case that I had not made, that the ban was a convenient lever to be reached for out of convenience.
I have not seen an operating system of civil governance that has not reserved to itself the right coerce behaviour as a "last resort" (this would be different things to different people), so it is hard to say much about your last paragraph.
I would say: if you don't think that people are going to debate and argue a case for banning, for punishing, for going to war or other forms of controversial and "extraordinary" behaviour, you probably can't think that people are going to agree to be educated properly, whatever that means. Debate, argument and loser's consent are fundamentals behind the model of democracy we have today. including, in my view, the part where the demos is educated.
Banning TikTok wouldn't take your free speech away, but the Restrict Act doesn't even directly ban Tiktok. It gives the government broad rights to censor any content it deems "undermining the democratic process" among other things, and gives it the power to enforce large fines and long prison time against people who help get around the ban (like providing a VPN).
The Restrict Act basically allows the US to impose Chinese-level censorship across the "Western" internet and persecute any one in the US who helps circumvent the censorship.
Sure it may be used to ban Tiktok, but what do you think will happen when the next whistleblower posts content online? What happens when the US finds cryptography for us normal people a threat to US democratic process? Ect ect...
Calling it the TikTok ban is just clever marketing sugar to make the censorship go down...
> The Restrict Act basically allows the US to impose Chinese-level censorship across the "Western" internet and persecute any one in the US who helps circumvent the censorship.
This statement is at least as misleading as the "Tiktok ban" rhetoric you're criticizing.
How? They could declare everything and anything as antidemocratic. There is legitimately no oversight to the Secretary of Commerce, which is an appointed position.
If the act is passed, the only thing stopping the government from taking it to such an extreme is the self-control of the government itself. That isn't very reassuring.
Websites that are critical to the governments message are blocked in the West too. Either blocking sites isn't only a symptom of a closed country or the West is more closed then you care to admit.
It’s mismanagement that arises out of European individualism. Because the individual is sacrosanct, you can’t infringe on individual “rights” like doing drugs or sleeping in parks. You can’t throw them in prison or execute the people facilitating antisocial behavior, like drug dealers. Simultaneously, no value is accorded to the collective right to a clean and orderly commons, free from antisocial behavior.
That just not true, there are a lot of places which value European individualism in the US or outside of it where drugs problems are on the same level as in Asian countries.
So what I'm seeing is Africa/Asia with the lowest drug use, continental europe with significantly higher drug use, the Anglo countries (which are more individualistic than continental Europe) even higher, and the U.S. (which carries western individualism to the greatest extreme) at the top of the charts. There seems to be something weird going on with the former Soviet bloc countries, but overall the chart seems to strongly support the trend I'm talking about.
Russia, Myanmar, Mongolia, UAE, Belarus are all significantly worse then almost all of the European countries (only Estonia is worse). I wouldn't consider those countries more individualistic than Germany for example. Vietnam and South Korea are at same levels as most of central/western Europe. India and Italy are also similar. There simply isn't any correlation (a formula if you will) between how individualistic a society is and how much drug abuse it has.
I actually think dysfunction in SF is not a result of European individualism. The situation is not best explained by individual vs collective rights. Rather, I think the dysfunction is deliberate and the politicians who control the city are part of a party and belief that repudiates anything European.
For example, the city of SF will come down on you like a ton of bricks if you are not from the correct demographic. There is a lawsuit where white police officers are suing the city for denying them promotions and a system of racial discrimination. What about their individual rights? Another example is how the city and state protect homeless encampments but make it difficult for regular people to build homes or other buildings. Or how homeless people and drug addicts attack and harass regular people and the city does nothing about it but it was up in arms when a former fire chief allegedly attacked a homeless person.
Individualism isn't broadly accepted, rather it seems the homeless have more rights than anyone else. This isn't individualism and shows that there is a ideology driving the dysfunction.
The amount of AI startups that are being restricted by this are minimal. Hugging Face isn't going to have to shut down because of this. University research into the best ways to stabilize GAN training isn't going to grind to a standstill. Companies developing weeding robots aren't going to have to turn to symbolic AI for image recognition. Scientists are still going to be able to model landslide risk caused by receding glaciers.
Maybe only 1% of AI start ups will be negatively effected by this to the point they would be better off leaving Europe.
I am not in a EU country and I would never ever vote to join the EU, but this is one of the better EU legislations. Saying this would "kill all startups" is like saying GDPR will kill all internet business in Europe. Its both alarmist and false.
The patriot act clearly compromises every 5-eyes nation. China and Russia have their own equivalents of a Patriot act, so hosting there is no good either. Mid tier and neutral countries can be compromised as in your example above. This is all truly depressing, but why does it have to be hosted on a central server.
I hate web3 hype as much as the rest of HN, but this seems like a genuinely useful application for it.
It played Stockfish 8, which was already a year old when they played, and the match was under arguably unbalanced circumstances.
The current version of Stockfish is Stockfish 15, and although it hasn't played AlphaZero (because AlphaZero doesn't compete), it has played clones of AlphaZero such as Leela Chess Zero in TCEC which it won against consistently for the past 5 seasons.
AlphaZero beating Stockfish was more a propaganda attempt from Google to appear as a world leader in AI then an actually attempt to find the superior engine.
Stockfish is definitely strong but the way LCZ or AZ beats its opponents is totally different and foreign to humans. Stockfish has a database and a whole bunch of openings prebaked into it.
Leela Chess Zero is definitely not as strong as Stockfish 14 but still ranked #4. That's not nothingburger.
> AlphaZero beating Stockfish was more a propaganda attempt from Google to appear as a world leader in AI then an actually attempt to find the superior engine.
Agree. I feel awful that Deepmind honchos take a round table meeting once in a while, where they pick up a board or online game and totally demolish the top player just to prove how good ML algorithms are.
And obviously consumer hardware is already being more optimized for running models locally.