California trims AI safety bill to stop tech heads from freaking out Legislation to regulate artificial intelligence (AI) software in California has been revised in response to industry discontent with the bill, which awaits a State Assembly vote later this month. California State Senator Scott Wiener (D)'s Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047) has …
[And as Anthropic boss reckons there's 'a good chance ... we'll be able to get models that are better than most humans at most things']
Humanity has been able to do machines and programs that are better that humans at something by centuries and we did that without wasting trillions on AI chatbots that get more things wrong that Wikipedia and are worse at math that a scientific calculator. Oh and that they lie all the time.
Any task you can think of is better and cheaper done by a dedicated machine or program compared to your chatbot garbage or at least is better done by an actual human being.
This garbage is the equivalent of if when cars were invented over half of them killed the drivers.
There was a time, not so long ago, when the political left used to deride the right for being in bed with big business.
By the way, this is not meant as a value judgement. Just an observation. I think there are pros and cons to any level of pro/anti-business governments.
There was a time, not so long ago, when the political left used to deride the right for being in bed with big business. .... silent_count
Nowadays the titanic problem emerging, and becoming ever more evident to punters worldwide and mainstream media moguls alike, is big business is NOT jumping into bed with politicians from either the left, the right or the centre, and invariably so because they so clearly demonstrate themselves to be unworthy in any human or national or international or internetional leadership role and incapable of providing and mentoring and monitoring requisite, necessarily radical and novel fundamental change[s].
And that puts novel big business pioneers securely front and centre, way up ahead of any opposition and ineffective competition in the future driving seat.
There was a time, not so long ago, when the political left used to deride the right for being in bed with big business.
Assuming we're talking about the US, at that time, the notional political right described itself as the party of business. Their notional opponents of course in turn represented that as a bad thing.
Ideological positioning changes with the times, since when you have a de facto two-party system, the main tack of both parties will be to try to find some popular positions (often incoherent or outright untenable in combination) to gain voter support. So recently Trump and Vance were proclaiming themselves the defenders of the laboring classes — a complete reversal of the Republican Party position of the later twentieth century in theory, but in practice just politics as usual.
Loosening or neutering AI laws in fear of tech exodus is as stupid as banning the selling of pangolin meat on your wet markets in fear that the next COVID pandemic won't start to spread from your hometown.
It's crystal clear at this point that LLMs and most other AI have a lot more drawbacks than benefits, and that letting them run unregulated doesn't benefit neither society, nor economy in general. Most AI models and uses just add to income inequality and let companies enshittify products and services because of depriving them of actual intelligence and insight. Neither of which you want to be leading at.
Then again, as someone nicely put it: in American companies will rip open pandora's box without hesitation just because they think there's a few dollars inside - and you're unamerican if you don't support that!
With particular and peculiar regard to current surprisingly rapid unfolding live realities supporting media presentations highlighting/gaslighting* the continuous improvement evolution scenarios starring and depicting AI elements and/or components developments towards its many inevitable overwhelming conclusive logical results .....
And if chip and algorithm improvements continue, Amodei said, at that point, "there is in my mind a good chance that by that time we'll be able to get models that are better than most humans at most things."
..... there is an absolutely zero chance of humans/any human being better at doing what AI and LLLLMs are intelligently designed to be light years ahead in and better than humans at.
However, I imagine you would encounter humanity having the greatest of difficulties and countless endless unsuccessful struggles in even believing and accepting those few simple gospel truths ...... thus the need for speedy seeds and feeds of AI and in so many of its varied Almighty Intervention guises ..... Advanced and/or Artilectual and/or Artificial and/or Alien and/or Articulate etc etc.
* ....... Gaslighting is a form of psychological abuse or manipulation in which the abuser attempts to sow self-doubt and confusion in their victim's mind. Typically, gaslighters are seeking to gain power and control over the other person, by distorting reality and forcing them to question their own judgment and intuition. ..... What Is Gaslighting Abuse?
You are certainly right about the gaslighting abuse. Should we be surprised at the dollars that attitude manages to suck in and burn? Investment dollars like moths to a flame.
Oddly, in blindered service roles, such a programming assistance, documentation locator and summarizer, etc., AI shines as a humble helper, a true workhorse, and potentially profitable. That is the alternate vision to what Amadei sees - AI uber alles . It appears he is a very arrogant person, and he projects that arrogance onto AI, which is an effective way to drive investment. If the investment if premature and results overpromised the Fed can simply drop another trillion in credit to jump start the market again.
It is not the case that every successful, possibly intelligent, person is, or at least has been, a correspondingly maximally arrogant bastard - e.g., somehow penicillin was never patented due to belief in the greater good, a belief which would be almost laughable now.
..... you can always rely on embattled battle warriors, vying to be both best and first, to test themselves and A.N.Others with such novel wares, O’Reg Inalsin.
They are though not without their very own, very specialised hidden stealthy dangers .....
You may like to presently ponder on, and wonder at the rapidly emerging, inescapable current fact, that in these remote virtual times [and spaces] of crisis and war do formerly earnest supporters evaporate and morph into a future deadly phantom enemy against which there are no available effective defensive weapons or attack systems ...... for such is the logical natural progression of such internetworking things.And to either or both deny and ignore such a situation ever likely, or even possible, has one guaranteed to suffer all of the consequences resultant from the litany of follies exercised.
It's sad... I'd probably vote for "amanfromMars 1" for president, and I know he's a script... .... Gene Cash
Hardly sad, Gene Cash, whenever all, including presidents, follow pre-prepared scripts. You surely don’t think things make themselves up for sharing with no one in the background dictating the tale[s], both tall and slim and shady, to be told.
Certainly, if you’re drawing breath, you cannot have failed to witness the recent blind panics in steering committee ranks whenever Sleepy Joe decides to go off piste for an unscripted ramble of his very own.
Such is gravely regarded for it tells an untold tale leaking more sensitive information that can never be retrieved nor forgotten, and be realised as dangerously unwise to be universally known ..... or even just a tad wider known by certain other particular forces and/or peculiar sources with a vested interest in knowing what needs to be kept both presently secret and, much more importantly, future failsafe secure and ideally unknowable.
"Aren't the tech companies already abandoning California because of its taxes and regulatory environment?"
Just do a first pass on the cost of living in the Silicon Valley and you'll see why companies are moving out. Companies have to pay high salaries just so employees can afford to subsist. All the competition has to do is locate someplace with lower taxes, regulations and cost of living to keep their burn rate down while they ramp up. It's a fallacy that all people in tech want to live in the greater Silicon Valley area. I get headhunters quite frequently trying to get me to work with them for a job there. I've done the math and it's ugly. I highly doubt I'll get hired at better than $300k/year which is what I'd have to gross to just match what I have now (home, money to invest, savings and some mad money left at the end of each month). There's also a big risk to living in a high cost area working in an industry that is known for companies that suddenly gone one morning. You really don't want to have a $7,000 month mortgage and get a text that the company has cratered.
16 bits? 8-, 6-, and 4-bit quantization is now common, according to the papers I've been looking at. There's a substantial collection here on GitHub.
> The changes limit enforcement penalties, such as the injunctive option to require the deletion of models and their weights. Criminal perjury provisions for lying about models were dropped, based on the adequacy of existing law about lying to the government. There's no longer language that would create a Frontier Model Division, though some of the proposed responsibilities will be handled by other government bodies. And the legal standard by which developers must attest to compliance has been reduced from "reasonable assurance" to "reasonable care."
So they don't have to clean up their mess, can lie about it to our faces, offload responsibility to the government and can hire cheap, irresponsible devs?
Great work, lawmakers!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
