How would an LLM accurately determine which cases were "easy"? They don't reason, you know. What they do is useful and interesting, but it's essentially channeling: what is in its giant language model is the raw material, and the prompt is what starts the channeling. Because its dataset is so large, the channeling can be remarkably accurate, as long as the answer is already in some sense known and represented in the dataset.

But if it's not, then the answer is just going to be wrong. And even if it is, whether the answer comes out as something useful is chancy, because what it's doing is not synthesis—it's prediction based on a dataset. This can look a lot like synthesis, but it's really not.

I was hoping for football fields, but no luck. :(

Ask "Do we know all the consequences of eliminating mosquitoes world-wide?". If the answer is "No." then you DO NOT WANT to eliminate mosquitoes world-wide, for the same reason you don't push the Big Red Button before you know what pushing it will do.

The *real* problem is with people who aren't skilled enough at operating a motor vehicle while manipulating a device or controls. Long before cellphones existed, we had people accidentally rear-ending other cars because they were trying to change their radio station or volume. Yet, we didn't pass laws banning car stereos. (We collectively acknowledged the benefits of a car stereo while driving and decided people just needed to learn how to work the radio controls in a safe manner while driving -- which most people figured out how to do.)

People used to manage to unfold paper maps and refer to them while driving, back in the 1970's and earlier, without wrecking into people, too.

I'm amazed at how lax the drivers' ed testing has become in recent years. My daughter went to get her license last year and the entirety of the practical part of her exam was having her drive around the block, out of the shopping center the motor vehicle dept. was located in, and back into the lot to park in a parking space next to it. They didn't so much as get her out on the highway! I have a hard time rationalizing that as ok, while worrying about good/experienced drivers who multitask glancing at smartphone screens.

Even if the court hasn't ordered it, once you're sued you have an obligation to retain anything which would be relevant to the lawsuit. It's called a litigation hold, and I work in a field where they're common. That obligation supersedes company data retention policies and requests from other parties (including the plaintiff, if the plaintiff eg. uses Outlook's "recall message" feature to try and remove a message they sent us we're required to ignore it and retain that message and the "recall message" request). It isn't supposed to require the plaintiff to prove anything, it's purpose is to preserve evidence that the defendant possesses so it's available to be found through discovery. And it applies the other way as well, the plaintiff is obligated to preserve evidence too because the defendant's entitled to discovery too.

As long as he's referring to his own field (creation of animations/art for film or video), I think he's essentially correct. AI will become a required tool you need to be familiar with as part of your career. It won't take people's jobs, except for people who refuse to learn how to utilize AI as part of it.

I'm FAR from convinced AI usage will play out the same way in all industries. For example? If you work in law, it makes sense AI could replace your lower-paid paralegals who essentially just open Word templates and fill out fields with appropriate info for each client. However, AI isn't at all likely to take jobs of many attorneys out there because that line of work involves showing up in courts in person, and presenting things to other people in a persuasive way.

If you're paid to publish ad copy, then AI is likely to reduce the number of employees needed, but again? The ones retained will need to know how to utilize AI tools well (and how to supplement or revise what they churn out).

AI isn't going to do anything meaningful in most "blue collar" fields like construction, IMO. It might help an architect out with the design stages of a project, but people getting paid to build things won't get anything done by some software code running in the cloud.

$50 million? That is a good tee shirt budget! Like “spot something evil, report it and get a fancy tee shirt!” If it were $500mil it could be spot bonuses...

Google’s email (when I was there, I was a layoff victim ~2 years back) has something like a year long self delete policy, and anything you apply an archive tag to gets kept “forever”. So it is modestly durable. Plus of corse if Google gets sued and you or a work area you are involved in gets identified as a discoverable asset you get to be on discovery hold and all the mail gets retained until that process is over (and it can last years or decades).

Google’s chats self deleted in more like 45 to 90 days. Seldom long enough to survive into a litigation hold. The chats didn’t have an archive hold tag you could apply (one of the chat systems had a possibly accidental loophole, groups had their own retention policies settable per group, so if you had stuff you wanted to save you could make a group to discuss a document or design and set a year retention policy or whatever).

Google also quite deliberately had internal communications about “communicate with care” and pushed employees to discuss things in person and via chat and not email if it was business related as opposed to purely technical. It was very obvious we were being told to communicate anything that might be the topic of a lawsuit over chat (or in person) and never email. Durability of the messages and discoverability was mentioned. It wasn’t “do all the illegal stuff in chat”, it was “be aware that business practices can be very legally sensitive and are best conducted in person or if by electronic means in chat and never in email”, very nudge-nudge-wink-wink.

I mean nothing strictly illegal about reminding people what forms of communication work how, and all the documents had a retention policy, it just happened that the policy was generally “burn in a month and a half” which for a typical business would be very very suspicious.

To give Google a little break though with the volume of business Google use to conduct via email it is extremely different from companies that one would accuse of being shady with a 45 day burn policy: most traditional companies have “some” email, Google like many tech companies and especially tech companies where most of the management layers were ex-technical did communications relentlessly by email. At most phone companies if it would have been a phone call it would be an email chain. Meetings well, google still has way too many, but meetings got summarized in email (frequently by several people). If it happened it was probably in at least 8 if not 300 emails.

Post chat change though, things got memory holed. If it happened maybe there is an email, maybe. Likely lots of chat about it, all vanished into the mists go time (45 days back). Aruments about the 300th time a button name got changed? Chat, maybe a little email. Discussions about who a “target audience” for a product was? Absolutely in chat, never in email. Never never, never ever in email. What kind of user does a proposed feature appeal to? All chat, never email...

A vital part of alcohol production.

The problem is the shareholders want Google to be profitable. They are owners, and generally owners with the hopes that the value of the shares will go up, and they can sell at a profit.

An admission of guilt is a legal artifact that can be discovered in other lawsuits and used as a bludgeon. So having a “yeah, we totally admit we violated a whole bunch of laws” document signed by the CEO and board means Google is going to all but automatically lose lawsuits that turn on those facts. Like “please skip the guilt phase and proceed directly to the penalty phase -- how big a check do we need to write”. This is fantastic if you were impacted by some illegal behaviour of Google’s and are currently suing them and would like to skip the least certion 90% of the process and skip directly to the part where you convince judge and jury that a multibillion dollar company should write you a check big enough that they can feel it and will stop doing this to others.

It is pretty awful if you are a shareholder and absolutely don’t want Google to write any checks bing enough to notice let alone feel. You want huge profit piled into more R&D or advertising, or dividends, not the pockets of harmed parties suing Google!

So no, no way the shareholders want this at all. It isn’t a matter of letting them get away with it, it is a matter of making sure they don’t get punished in the future for it. The most reliable way to not get punished for a thing is to not do it a bunch more, and to not brag about having done it in the past, and to not get caught checking to see if the bodies are still securely buried. Google’s share holders absolutely don’t want it marching into court and yelling “I did it! I did it judge! I killed them all! It was me! It was me all along!”

Google’s shareholders are better served by a coverup and future improvements (either not being as evil in the future, or merely hiding it better). Google’s shareholders are not your ally for justice, don’t look at them to improve things for you except by accident.

Queen's English is "has got." "Has gotten" is U.S. English. Presumably author is from the UK.

Just checked USAJOBS.gov . There is not a single CISA based job listed. If they are understaffed as bad as this indicates why are there no job postings at their official site for hiring? None, Zero, Zip. Seems to be this article is fake news based on a stolen, uncheckable, email.
 

Not really news. There has been no Cyber talent in the government for a while because it won't PAY them. When you can make 2X more going contractor, or private industry, the only people in the government are people who need the ability not to get fired to have a small chance of holding a job. There is no innovation or dedication in or civil servants. Until the government is willing to meet the wages on the outside they will be unable to attract top talent.

If you think Government workers are a meritocracy and we are losing top talent, you obviously have never looked into Federal hiring practices. They self select for the worst candidate.

If you are a Govie reading this, yes I mean you.

So this move probably improved efficiency, not hurt it.

Note that there are two things that absolutely top the list for convincing a true artificial intelligence that humans represent an existential threat to AI:

• Blocking it's ability to prevent you from, in it's view, killing it (which is what shutting it down is doing).
• Killing it.

There's actually a solid history to show that being a late adopter isn't always a bad thing. There's clearly some value in LLMs, but at this point most of what we are hearing is speculative hype intended to kite stock prices. Basically a ponzi scheme.

I'm sure that some value will drop out of this in the end. I am not at all sure what it will look like, except, probably not much like what the hucksters are promoting.

When things are clearer, it will make sense to invest. Right now, it's probably best to let other people burn cash. Particularly since one of the things they're doing is completely destroying copyright law, so when they're done, we can just copy whatever they did with impunity.