Re: URL for Pick
Here's a Dick Pick for you:
https://m.youtube.com/watch?v=9Sz3MDl1XRw&t=38s&pp=2AEmkAIB
Poor sod apparently died at the age of 56.
552 publicly visible posts • joined 19 Oct 2021
Damage limitation:
1) Disallow email addresses in the CC field that have a different domain than the sending address. This could be enforced by the MTA rather than the client.
2) Client shows a warning dialog requiring confirmation whenever attempting to send an email with more than 1 address in the CC field. The default action is cancel.
Hmm, which bit don't you agree with?
That, for example, Lehman went bankrupt due to the 2008 crisis and Goldman Sachs didn't (and indeed made a profit by betting against mortgages, see https://en.wikipedia.org/wiki/Goldman_Sachs_controversies#2008_financial_crisis ) is pretty objective.
Specifically to Lehman, I'll add this anecdotal story out of interest: https://www.informath.org/apprise/a8800.htm
I only mentioned Lehman as an example, and Golman Sachs as the exception in that they apparently listened to their lowly risk analysts,
While the US Government did play a minor role in the cause - relaxing lending rules in the *1990s', and repealing Glass–Steagall - it was mainly greed and stupidity by banks (esp. synthetic CDOs) that caused the crash.
That no-one responsible went to jail (let alone was strung from the nearest lamppost) means the next crisis is pre-destined.
There were risk analysts (often with an engineering background) at some investment banks who saw the danger at least a year before the crash.
They were, unsurprisingly, mostly ignored by their higher-ups. I believe Goldman Sachs may have been the exception, hence they're still around and firms like Lehman Bros. aren't.
Leopold Aschenbrenner also worked for OpenAI (until fired) after FTX. That sort of person.
His book was completely invested in "AI" related stocks. No diversity.
His only "hedging" was shorting software companies - so basically exactly same bet as his long positions.
He was hugely leveraged - apparently 4x.
He didn't have liquidity in reserve to cover his leverage if his stocks dropped temporarily, hence the margin calls and the firesale.
Hardly a "good stock picker". It's easy to make a plus with leverage in a bull market...
The wolves of Wall Street saw this overconfident amateur with too much money (mostly other peoples of course) and took him to the cleaners at the first opportunity.
Took 3 weeks to get from Germany to the UK
That'll be one of those Brexit benefits we've heard about ;-)
Fountain pens are one thing I love getting from AliExpress
Fair enough. With the observation that a faulty fountain pen will at worst ruin a shirt; a faulty charging device will potentially burn your house down.
The idea that it's important, for say the EU, to have sovereign "AI" is implicitly buying in to the idea (or hype or lie if you will) that what's currently being sold as "AI" is actually useful and important.
It isn't. Currently non-measurable ROI for most customers, "hallucinations" (i.e. errors) unavoidable, actual real costs not going to justify any possible benefits in most use-cases.
Local (relatively speaking) LLMs specialised for a certain use-case (e.g. aiding medical research) running on not-so expensive and heat-producing hardware as now is the likely future.
From the article:
working in the Children and Young People directorate
Arguably, one would imagine that certain roles in this department are justified in being able to answer questions like "has this child had an usually high amount of injuries?" (medical records) and "do the the parents have a criminal record for violence related offences?" (criminal records).
Obviously this access needs to be tightly controlled and audited. Though given that the idiot in the article was caught (and the impression is fairly quickly), at least part of the system is working.
The conspiracy minded would say that that's the plan with "AI" (aside from mass unemployment) : make the population even less capable of critical thinking than they already are.
Icon: obvious.
For the record, no, I'm not being serious. Although I could imagine some nefarious people seeing it as benefit.
Not used Haskel but I imagine it's advantages such as strong type safety mean compilation takes longer than with other languages. You don't get owt for nowt.
Using a faster development iteration language may be legitimate if that's the critical factor, but one needs to be aware of the then corresponding downsides of doing so (often runtime errors instead of compile time errors).
Moving a project to another language just so it's easier to use with so-called "AI" seems like a decision where I hope the maker has an organ donor card (as they appear to be brain-dead).
I beg to differ.
Language Models are dependent on their training data, which has been created by *other people*, and in many cases often without their permission, under licenses which limit what may be done with their work.
If I translate a book* while I may be the automatic copyright holder of the translation (as I've put a significant amount of effort in myself), I can not publish said translation nor give someone else permission to publish it (a license) without the permission of the copyright holder of the original work. If I've used a machine to do the translation with no input myself, then I'm not even the author of the translation.
Entering a prompt to a LLM is not "sufficient human control over the expressive elements" to be the copyright holder of the output.
So while LLMs may be considered a tool (though more so in the British vernacular sense), they are not the same as other tools in this aspect.
You can't for example copypasta BSD Licensed code from various sources, then release the result under the GPL. Using an intermediary such as a language model to do the same doesn't free you from the legal or (IMO more importantly) ethical considerations.
-----
* one of those things made from dead trees
This is a good point.
Only the copyright holder can issue a license (e.g. GPL).
When the code comes from a LLM, who is the author (per Bern Convention) ?
It's not the meatbag entering a prompt (precedent long set about the significant own creative input required) . The output is at best a derivative work.
The output of a LLM can only be GPL'd is you're sure the training data used to generate the output was GPL code (which you can't be). One bit of say BSD-licensed code used to generate the output and you can't license it under the GPL.
In this case, it is GlobalFoundries (GF), an international business with its headquarters in the US...
The pair say that GF's Dresden fab is establishing a European manufacturing flow with funding from the European Chips Act. This will ensure that every step of the production process occurs within the EU, so that no sensitive design data leaves the region.
(Emphasis mine).
If you believe that last statement, I've a bridge you might be interested in purchasing.
The design will be sent/available to head-office in the US. There's also a risk of a TLA "insisting" on design changes at the factory level.
Now admittedly, even if it was 100% sovereign there's still nothing to stop a state actor at least attempting to bribe and/or extort someone at the factory to do the same, but it would be a level of more difficulty.
Given that Samsung sells lots of other things as well as memory to "AI" companies, suggests their mark-up on RAM is even higher (assuming the mark-up on other items is till around the 6% mark - if some sections of the business, e.g. phones, are making a loss at the moment, that would push the RAM profit rate even higher).
Misanthropic and OpenAI need to go public to raise more capital that they're burning through at a rate of knots.
On the other hand, going public will subject them to more financial scrutiny than they are comfortable with, with the risk they don't receive the desired valuation.
The only way they're not going to deflate, if not actually pop, is if they're classed as Too Big to Fail* and are bailed out by the US government's deep pockets, directly or indirectly (E.g. the Pentagon is made to spend $500B on "AI" for example. The recent "Trump Accounts" where money from the state will go into US based indexes, boosting stock prices, is a related version).
Oracle should by all rights implode by 2030, except its owner's friendship with the current (and 50% chance future) administration may save Larry's yacht collection.
---
* If it wasn't for the current tulip planting craze boosting the GDP figure I suspect the US would be in a technical recession now
Legislate that for software, such as many games, that requires a propriatary online service to order to function, the service must be provided for a minimum time since last sold as new (not the release date).
Alternatively, when the service is a licensing system or similar, releasing an updated version that no longer requires said service within the minimum support time is also acceptable.
I'd suggest 7 years as a reasonable time (also matches the new EU spare-parts rule for phones & tablets).
From the article:
pointing out that Texas had 14 GW of new capacity announced
Texas is, as far as I know, most of the time quite warm, which is usually suboptimal for data centers. Somewhere like Montana would be more intuitive.
Secondly, the electricity grid in Texas is well-known to be pretty fragile. Or are they planing on burning directly oil coming out of a nearby hole in the ground to power the damm things?
The EU doesn't appear to have more "no-gun zones" than the USA
In most European countries *everywhere* outside specific areas for hunting (controlled, time-limited, regulated, etc.) or military training grounds (military only obv.) or shooting-range is a no-gun zone, unless you're a police officer.
I'll concur that at least part of the US's problem is cultural (and not just with guns...).
The UK was already pretty restrictive for gun ownership. Then in 1996 we had the Dunblane massacre, and pretty much the whole country went "fuck that, we're not risking that horror happening again" and gun laws were further tightened with widespread support, so much so that biathlon athletes often have to do half their training abroad these days.
IINM, Australia was somewhat similar with Port Arthur.
The problem with the US, that despite having experienced hundreds of the equivalent of Dunblane, it's not done much about it due to the US's mythological Individualist culture and so many yanks simply don't give a shit about other people .