Posts by Michael Wojcik 12244 publicly visible posts • joined 21 Dec 2007
My laptop does too, and it has a much better (if still kind of rubbish) keyboard. And it'll do 80x70 text. In fact it'll do two of them side-by-side, with some real estate left over.
I can see use cases where "virtual ssh device" could be handy, but they're mostly "someone did a shit layout of the machine room" – compensating for a fundamental error. I prefer to solve that problem by never being in the machine room in the first place.
Nah. At this level, the salary generally isn't important.
What you get, mostly, is connections. Sure, there are the benefits, and there are the stock options (though in Twitter's case those will probably soon be worthless except to collectors of ephemera). What you want are the lucrative, minimal-effort speaking honoraria and short-term consultation gigs and maybe the management-book deal (those things get pumped out by the zillions), and most of all the board appointments. Getting five or six figures for four meetings a year is a pretty good deal, and once you make those connections you can be appointed to half a dozen of the things.
As I wrote when her appointment was first announced, I think it makes sense for someone who's interested in building their own reputation and doesn't really care about anything beyond that. She's being "gutsy" by taking on a company that everyone knows is in dire straits, under the leadership of someone who's clearly difficult to work with. Because Twitter is widely seen as doomed, either she succeeds against great odds and is brilliant, or she fails but it's not her fault – and along the way she gets to show off making "tough choices" (tough for other people, that is).
Executives who leave failing companies rarely seem to be held accountable for those failures. Indeed, high-profile failures appear to help their careers. It's good work if you can get it.
music and culture also has an incredibly short shelf life
Right, which is why no one these days has ever heard of Star Wars, or Shakespeare, or Journey to the West. And no one listens to Bach or traditional folk music.
Hell, my 10-year-old granddaughter spends her free time reading books and playing games that are older than her (she and her friends are back into Minecraft now, which came out in 2011), and I've heard her singing songs from the 1970s. The six-year-old was obsessively re-watching Disney's Snow White (1937) not long ago (in rotation with Encanto, from two years ago, and Zootopia from seven).
I'm in favor of sharply reducing copyright terms myself, but I don't think the "no one pays attention to old stuff" generalization holds any water.
How did they "corral the commons"? Your work is still available. Their derivative work isn't, to the public, but it never was. The situation is unchanged.
If you don't want work derived from your work1 to be proprietary, fine; that's your preference, and I'm glad you have a legal mechanism that might actually achieve that.2 But I've never been persuaded by the FSF's claim that this is somehow a moral imperative (and I was in Cambridge not long after the FSF was founded and there was still much contentious debate over it, and Stallman's pronouncements, and a few years later over the GPL). There is quite a lot of finger-wagging done by GPL advocates but not a great deal of actual substantive argument to justify it, beyond personal preference.
1Of course all labor that's not the most primitive sort of extraction is derived from prior labor. Actually even primitive extraction is, since adult humans do not spring forth fully formed from the void, generally speaking.
2In practice I doubt the GPL has been terribly successful at this. Certainly we've seen no shortage of stories of "vendor X is shipping GPL'd component Y with changes but not providing source".
Twitter was "the most important platform in history" and "immeasurably valuable,"
Plenty of people never believed this bullshit. I didn't, and neither did many Reg regulars.
person we previously all worshipped
"we ... all" who? Again, plenty of folks here calling him out for being an arrogant ass long before the Twitter debacle.
Now that he's paid for it, Musk is dumb, an idiot, a fool
True before, true now.
naturally everyone's still using it
No, not everyone.
You seem to have a rather dramatic inability to perceive reality.
Originally, as applied to stocks, it meant simply high-priced shares. One example cited by the OED:
1874 San Francisco Chron. 5 Jan. 2/2 If times are good and the market flourishing, the game may be played with ‘blue chips’, as a gambler would say, the very high-priced stocks being the favourites.
Later it was frequently used to mean large-cap stocks for firms that had been around long enough to have weathered at least one downturn, and often to mean stocks that reliably paid dividends and so were good long-term investments.
I don't think Twitter ever satisfied any of those definitions, so your point stands, I think.
I strongly doubt any of the "advice" quoted in the article would be useful for any overeating condition. It's not like people don't understand the relationship between consumption and weight gain. Obesity has a wide range of causes, and – from what I've read – calorie-counting and body-measuring are rarely helpful, at least in isolation. (Counting calories is particularly misleading because human metabolic efficiency is very sensitive to diet and activity. People aren't furnaces.)
I have sympathy for people who find it difficult to achieve a healthy weight (whatever that might be). I've stayed more or less the same weight my whole adult life, but that's certainly not because of any virtue I can claim. I don't police my diet significantly and I don't exercise for the sake of exercise. There are no doubt various hereditary, environmental, and economic factors at work, but no "willpower" or "discipline" or "smart lifestyle choices". It's just luck, and I could have just as easily found myself on the other side of that coin flip.
The role of judge differs considerably across jurisdictions, so as a generalization "AI as a judge" doesn't mean much. Of course, "AI" is also a largely meaningless term; unidirectional transformer models would be terrible judges for all but the simplest and most straightforward of cases, but there are many other possible architectures. Not that I'd want any of them presiding over my case, at least at the present state of the art, and even though many human judges are awful.
Steinkamp said "this basic function should never have been able to be executed if the owners of the asset had run basic security and hardening efforts prior to releasing it into the production environment."
I'm calling bullshit on that. A quick review of Molly White's site shows any number of cases where bugs in "smart contracts" were exploited despite those contracts having been audited by one, or often multiple, security firms that claim expertise in the area.
No doubt the Jimbo team were lax in basic security; I don't find that hard to be believe at all. But apparently finding all vulnerabilities in smart contracts is a very difficult problem in general, and even (self-appointed) experts are pretty bad at it. This vulnerability might have been caught, but very likely others would not have been.
I don't think the whole DeFi / smart contract approach is salvageable. Which is no loss, in my opinion.
It's not really terribly complicated; it's just that because there are so many flavors of exploits against DeFi, each story tends to be packed with technical details that aren't actually important if you don't care about them. Basically:
1. Someone borrowed a bunch of (notional) money
2. They bought a bunch of X
3. That drove the price of X up
4. They sold the X back at the inflated price. Here there was some fiddling with a vulnerability in the system to permit this action.
5. They paid back the loan from step 1 and kept the rest as profit
The only subtleties are in the fiddly bits of step 4, and the fact that all five steps are completed in a very short period of time.
"meant to be" by whom? The Athenians, who extended the franchise to citizens, who all had to be free adult men – probably less than a third of the adult population? Yeah, that was great. Or the US "founding fathers", who had a similarly restrictive idea about who should be given the vote, and also greatly limited what was subject to voting in the first place?
And that's just in the US, of course.
Eternal September is eternal.
While routine use of agricultural antibiotics should be banned – and many countries have done so – overuse by humans is at least as big a problem. From what I've read, there actually isn't a lot of evidence of antibiotic resistance in livestock transferring to common human pathogens, but there's plenty for developing antibiotic resistance in pathogens already in people.
Resistance can also arise when human patients receive antibiotics for viral infections or when they do not otherwise need them.
Very true. I was just reading an old Derek Lowe post where he cited a source stating that a large majority of antibiotic prescriptions in the US were for non-bacterial conditions. Patients demand them, and it's not worth the liability for doctors to refuse them. It's a big problem.
Lowe suggests putting more research into antibiotics that target host pathways the pathogens use, since the selection pressures for overcoming those are lower. A pathogen that evolves to target a different pathway will eventually outcompete ones that don't and those that pick up the change (through the various mechanisms by which prokaryotes distribute genetic material) will be more successful at reproduction, but the vulnerable ones won't be killed off immediately.
That said, new antibiotics of whatever sort are welcome, particularly when research on them has diminished because they're much less profitable than drugs which patients take routinely for years (e.g. statins, anticoagulants).
it's stranger than we can imagine
This is in a sense trivially true: just compare the number of possible discrete states of the human brain1 with the number of possible states of the universe. There will be many more states even of the visible universe2 than of even all human brains combined, so we can't imagine all of them.
1At the level of resolution where change is significant, e.g. some synapse does or doesn't fire.
2And based on WMAP and other experiments, apparently much, much more outside our Hubble volume, but of course that's forever inaccessible to us (assuming causality is preserved3) and so might as well not exist for our purposes.
3And if it isn't, things will definitely be stranger than we can imagine.
Well, they got the initial law passed, and by exempting agricultural equipment they avoided a huge lobbying blitz by Deere. Now a few years later as R-to-R gets established in other states and civilization doesn't collapse from it, they'll have ammunition to push back when they amend the law to cover agricultural equipment as well. Or Deere and the other tractor OEMs will see this is a Prisoner's Dilemma and one or more will defect and open their stuff, for the PR and sales benefits; they must see that the writing is on the wall.
X86 has come closer than anything else in history
Has it? Last I looked, there were still a lot of Z80 cores inside little dedicated-purpose embedded chips, for things like kitchen-appliance controls. That era may have passed now too, but it was still the case circa 2010, which would give the 8080 architecture a good 35+ years. Wikipedia says there's still a Z80 variant in TI calculators as of 2015. True, x86 is only 5 years younger, so it won't take x86 long to surpass 8080 after the last version of the latter ships, but for the moment I believe it's still telling its 16/32/64-bit younger sibling to get off its lawn.
(Of course you could pretend that 8080 is part of the x86 family, if you really wanted to loosen your definitions. And you could pretend the ROMP is part of the POWER family, which makes it more or less contemporary with x86. In any case, 360 beats them all handily, with a solid decade on 8080.)
IA-64 wasn't a complete failure in the beginning because of marketing. It was a lousy CPU which was dreadful to use, particularly if you were trying to write a compiler backend for it, or (god help you) debug something in assembly.
Even the things that seemed like a good idea in theory – hey, a trap representation for integer registers! – were terrible in practice because they weren't handled well by the OSes that ran on IA-64. Which was mostly HP-UX.
We're likely not far away from the point where many consumers let algorithms do a lot of their discretionary shopping for them. Already we have plenty of popular "sell me stuff without making me choose" services, like those meal-preparation ones or Bespoke Post. I know people who have gotten random-clothing-of-the-month subscriptions and the like.
It's a small leap to just turning a budget over to a purchasing app that "learns your preferences". Or even not budgeting it, if it's reasonably successful at matching how much the victim consumer would have spent on unnecessary crap every time unit anyway. I wouldn't do it, but I'm a contrarian curmudgeon and I hate all joyful things such as spending money. You know there's no shortage of people who would. Every day is Christmas!
Then those machine-generated ads will gradually shift to targeting purchasing algorithms, rather than human purchasers, and we'll have an arms race between sell-bots and buy-bots. Which, again, should not be surprising; that's what we've seen happen in financial trading, for example, and in IT security. We can just take humans right out of the loop, and simply spend a lot of resources making and shipping crap around the world to be briefly admired and then tossed in a corner.
O brave new world (same as the old world).
Exactly. The expected return from advertising to a consumer something they already frequently buy is minimal, unless you think you can tempt them away from a competitor. Well, there's the theory about "brand prominence", which is what supposedly justifies the excessive advertising of things like Coca-Cola and F-150s here in the US; but I'm dubious about its value, frankly.
Yes, it's just possible there's some imbalance of power potentially affecting the camera owner's decision to permit or refuse access.
There is a reason for the principle of requiring warrants for searches. That reason does not magically go away when the police promise to be nice and a craven city government rubber-stamps their requests.
From the article: a 100,000 qubit "quantum-centric supercomputer" allegedly capable of solving the world's most intractable problems
Did IBM actually claim that? Because it's prima facie bullshit.
There are only a handful of areas in which we know of algorithms that have a quantum advantage – i.e. are in complexity class BQP (or possibly some related classes; see the Complexity Zoo). They most certainly do not cover most of "the world's most intractable problems".
Err ... we have quantum computing today. We might have quantum advantage today, e.g. in Google's quantum-circuit experiments.
If you had written "useful quantum computing to be 10 years away..." you might have been right, though there's a decent chance that one or another of the systems under development will eventually scale up to a number of error-corrected qubits that is actually useful for physical-simulation tasks.
Shareholders have no control, but they can sell shares, depressing the stock price, reducing equity and raising the cost of borrowing. A significant shareholder revolt could hurt Meta, in principle.
In practice, there won't be one. Institutional investors mostly move slowly, and Meta have the capital to pacify them with buybacks (or dividends, in theory, but dividends are passé). Unmanaged funds, such as index funds, that hold Meta shares buy and sell according to their rules, not sentiment. Big activist investors generally retain their holdings in the hope of having influence.
There are certainly many good applications for Merkle graphs in general, such as some types of filesystems, git,1 some database systems, and so on. And there are applications which are of theoretical interest even if they're not practical, such as hash-based signature schemes.
But blockchain is just a dumbed-down, degenerate Merkle graph, and I've yet to see a good use case for it. The one where it supposedly provides something useful – as an append-only ledger in a Byzantine environment – fails all over the place in practice (e.g. partitioning attacks) while consuming egregious amounts of resources.
1Well, git is a popular application of a Merkle tree. I'll leave the question of whether it's a good one to the side.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
