* Posts by PghMike

123 publicly visible posts • joined 28 Feb 2011

Page:

ChatGPT starts spouting nonsense in 'unexpected responses' shocker

PghMike

Probably finally ingested the lyrics to Close to the Edge

My guess is that Chat GPT finally got around to ingesting Jon Anderson's lyrics on Close to the Edge.

:-)

Forgetting the history of Unix is coding us into a corner

PghMike

MacOS Unix server?

I don't think MacOS has a Unix server. I'm pretty sure that the Unix system calls are supported directly by the OS kernel these days.

I certainly don't see one running on my MacBook.

Air Canada must pay damages after chatbot lies to grieving passenger about discount

PghMike

Air Canada must really be terrible

I can't believe they took this case to court instead of just paying the $1000 or so difference, esp. considering their pathetic justification that the chatbot was its own entity. What idiots.

FWIW, we had our own problems with Air Canada. We purchased a refundable ticket and then canceled (well within the cancelation period) a flight with Air Canada. They refunded everything but a $46 seat purchase fee, which for some reason they claim we have to recover from their partner airline, who of course refused to refund money they never received.

I thought Canadians were supposed to be reasonable, but AC sounds like it's run by dolts.

TSMC finds its green chips are highly sought after... the edible ones

PghMike

Where do you get the newspaper?

They're almost all gone here in the US.

War of the workstations: How the lowest bidders shaped today's tech landscape

PghMike

Re: A lot of design points were being explored at that time

They also had absolutely no security. If you put one of them on today's Internet, they'd be hijacked for crypto mining in 3 minutes.

PghMike

You must be kidding

*Everything* in this article is wrong.

My MacBook is much easier to use for a normal person than the systems I used at MIT in the 1970s. If you told someone what it was like to use MIT's ITS (Incompatible Timesharing System), they'd think I'm making a (stupid) joke. To login as "ota" you'd type "ota$u" -- no password required. And then you were logged into the DDT debugger instead of a shell. The whole thing was written in PDP-10 assembly.

The differences between these languages is invisible to me -- Dylan looks like virtually every object oriented language I've ever seen; it's hardly amazing.

"Rich file systems, with built-in version tracking, because hard disks cost as much as cars: gone. Clustering, enabling a handful of machines costing hundreds of thousands to work as a seamless whole? Not needed, gone. Rich built-in groupware, enabling teams to cooperate and work on shared documents? Forgotten. Plain-text email was enough."

*Seamless clustering* in the 1980s? What planet are you talking about?

Rich built-in groupware? I have no idea what you could be talking about. Certainly Xerox Altos didn't have that, and they were the most advanced systems from back then. PDP-10s didn't even have a GUI, and the less said about the user experience on OS/360 the better.

People loved Lisp machines because they were some of the first systems that had a GUI at all. You had to program them in Lisp, which, in case you missed it, was a pretty unreadable language. It had some clever parameter binding rules that let you create closures, but otherwise it was a pretty low level language. C++ typing is much easier to use. Modern programming languages exist to help large teams to work on the same projects. You need clean interfaces with enforced module isolation and modern C++ versions exist to do just that.

Editing the operating system you're running as you're running it -- what could possibly go wrong?

"Firstly, because the layers are not sealed off: higher-level languages are usually implemented in lower-level ones, and vulnerabilities in those permeate the stack."

No, sorry, C++ compilers are written in C++. I haven't heard of a high level language compiler written in assembly for decades.

You want some good system design advice from that era? Here, read Butler Lampson's Hints for System Design. It's on the web here https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/acrobat-17.pdf Lampson seems to think that virtual memory is much harder to deal with than it is, but aside from that, his advice has aged well.

From Joaquin Phoenix to Rowan Atkinson, we enjoyed your Musk movie casting calls

PghMike

Christopher Walken

I'd vote for Christopher Walken to play Musk. Walken does 'psycho' pretty well, and the odd speaking cadence of his characters matches Musk's to a T.

But Kermit the Frog also would be good -- odd speaking cadence, inability to make eye contact (no eyes!), same thin lipped smile.

My wife says Walken's too old, though I think Musk looks a lot older than he is. She suggested Adam Driver, who also can deliver lines with a quirky cadence.

Why Chromebooks are the new immortals of tech

PghMike

MacBook Pro doing fine

My 2016 MacBook Pro (16GB, 512 GB SSD) is still doing fine, here in 2023. My only complaint is that it doesn't run the latest MacOS, which prevents me from running the latest Xcode, which fixes a bug preventing ti from connecting to my iPhone running the latest iOS. Everything else works perfectly, the machine is still zippy, and it runs the real MS Office, which is unfortunately necessary in the real world.

I'm really looking at the 15" MacBook Air, since it is 10 ounces lighter, but I'm too cheap to do that while this is still going strong.

GNU turns 40: Stallman's baby still not ready for prime time, but hey, there's cake

PghMike

RMS contribution

I haven't really even talked to RMS since the late 1980s, when he was bopping around CMU getting misled by the Mach 3.X people about how efficient their microkernel architecture was: Sys calls were really slow, but once you mapped a file, reading the next byte was fast, so they'd tell you how fast 'getc' was, since it was basically *data_pointer++; in reality, everything else was slow. IIRC they also convinced IBM to structure (a version of?) OS/2 around the same micro-kernel principles, with the result that we all know.

But despite Hurd never coming to fruition, everything else he did, including inspiring the entire Free Software movement, has been a massive success, well beyond what I, or probably anyone else who knew him back in the 1970s, could have believed possible.

My vague recollection from the late 1970s, was that he hated AT&T for refusing to make Unix(TM) free, and he decided that in retaliation, he was going to rewrite everything they did and give it away. I mean, it was obviously ridiculous -- he's just some guy (TM), and AT&T was The Death Star.

But looking back, of course, he actually did it. Linux provided the OS, of course, but the compilers, editors, and more tools than I can count all came from GNU/FSF, and no small number of them were written by RMS himself. I've known my share of awesome programmers in my day, but i can't think of anyone who cranked out as much well designed, if poorly documented, code, than RMS. Not even close. The Free Software model has taken over the world, leaving AT&T in the dust, and even convincing Microsoft to join rather than fight.

So, let's all raise a cold one in Stallman's honor.

PghMike

Re: 1991

Exactly -- Linux was just a cool toy in its first release.

The Anti Defamation League is Musk's latest excuse for Twitter's tanking ad revenue

PghMike

Elon vs. ADL

Well, I'm pretty sure that in the US, a public organization like Xitter can't really sue anyone for making true statements about it.

But if Lonnie really wants to give it a go, I'm sure the optics of suing the ADL will *really* help improve ad revenue from already skittish advertisers. I'd guess that already just his *blaming* the ADL for his moronic operating of Xitter is already going to scare more advertisers away.

Musk's X tries to win advertisers back with discounts

PghMike

Xitter?

I'd lose the W, and just call it Xitter. And then use the Pinyin pronunciation for it. Just like Xi is pronounced "Shi"....

EU's Cyber Resilience Act contains a poison pill for open source developers

PghMike

Apple & Open Source

Doesn't really change the thrust of the article, but Apple most certainly uses open source software. Just go to the General / ... / Licenses part of the Settings app, and you'll more screens than you can count of FOSS licenses.

They may maintain a walled garden, but the mortar holding it together is open source.

ChatGPT becomes ChatRepair to automate bug fixing for less

PghMike

That's just great, Ollie

A way of generating bug patches that doesn't require understanding the underlying problem. What could go wrong?

Turing Award goes to Robert Metcalfe, co-inventor of the Ethernet

PghMike

Don't blame Metcalf for CAT 5 plug

Way back in the day at CMU, I actually used the original 3 Mbit Ethernet, due to CMU's having received a grant from Xerox PARC of a dozen or so Altos, a Dover laser printer and a bunch of Ethernet cards. My recollection was that the Ethernet was just a long Cable TV cable, and each machine on it would tap into by drilling a hole through the outside of the cable until the core wire was exposed. You then inserted a "tap" which was just a probe that touched the central wire. Adding a tap, however, often caused signal reflections around that spot.

All those reflections from new taps meant that adding new machines could disrupt the signal for other connected machines on the cable. Adding new machines to a cable was basically an art since doing anything meant that every pair of machines on the cable had to be rechecked to make sure the reflections didn't mess up communications between those two machines.

Point to point cabling was a welcome and important improvement over this. OTOH, the single cable approach worked better than you'd have thought. CMU's Andrew project started using the single cable approach.

Google's AI search bot Bard makes $120b error on day one

PghMike

Re: Top of hype reached, I'd say...

Pretty good summary -- these things are pattern matchers, and surprisingly good ones when fed enough data. But they're also quite fragile when fed less than immense amounts of data, as you can never tell exactly what features their recognizers have glommed onto. In all likelihood, you will be surprised sometime in the future when they fail on something obvious.

ChatGPT is worse. It looks smart, but it doesn't really have a good model for the world, so it often ends up just making stuff up. And it doesn't even know that's what it's doing. Bummer that Google's demo screwed up visibly, but really, this stuff isn't ready for prime time with either Google's or Microsoft's technology. Not even close.

India's Supreme Court finds Google's appeal against monopoly fines unappealing

PghMike

When does Apple get examined?

Apple's behavior is generally even worse than Google's in controlling what apps run on its devices. It'd be nice to see Apple forced to open things up as well.

Disruptive innovation's like a party. It's always happening elsewhere

PghMike

Innovation is always dying

It always looks like innovation is dying, because you can’t see what was truly innovative until 40-50 years have passed. When I started programming in ‘72, no one thought computers would revolutionize anything — they were just giant adding machines. I wrote a program to compute e**x, and was very proud of it, but I couldn’t see how pervasive computing would change the world.

There were those brighter than I that did make those predictions, but those predictions didn’t show up for decades.

Today, it is obvious that we’re at the cusp of all sorts of biological and drug creation innovation, which I’d guess will be as big a deal as computation over the next 50 years. Learning how to harness and scale ML will likely also be important. But we can only glimpse the possibilities at present.

Too big to live, too loved to die: Big Tech's billion dollar curse of the free

PghMike

Other revenue opportunties

Obv. don't know how G monetizes things, but there's clearly an advantage to having 1.8 billion people who know all the ins and outs of your mail application. If nothing else, it probably provides a pipeline of companies ready to pay the $72/user/year for the minimum product.

There are apparently about 7M companies using this stuff, and even if the average company is only 10 people, and they're all on the minimum priced plan, you're talking about $5B in revenue, and those are probably very conservative estimates. That's not trivial for a company with $64B in annual revenues. And its not clear they could compete at all with MSFT's MS 365 suite without that pipeline.

Crypto craziness craps out – and about time too

PghMike

Re: Blockchain next..

Exactly. Blockchain is the world's worst implementation of a database; it would be a struggle to come up with a less efficient implementation that wasn't deliberately designed to suck.

PghMike

Binance.US a mere front?

This guy thinks that Binance.US is just a front for Binance, to get around the fact that the latter isn't allowed to operate in the US.

https://dirtybubblemedia.substack.com/p/is-binanceus-a-fake-exchange

It's pretty clear that the only application for crypto is money laundering. And that's not going to age well as the DOJ comes after the crypto exchanges.

FTX's crypto villain Sam Bankman-Fried admits 'I made a lot of mistakes'

PghMike

Smells like teen fraud

Sure looks like SBF took FTX deposits and without permission of the owners, lent them to Alameda with collateral consisting only of FTT tokens whose price was BS. That is, FTT token prices were set based on trades between FTX and Alameda, both of which he controlled.

Pretty blatant fraud.

Too soon? Amazon commissions FTX mini-series

PghMike

audits for VCs?

Can anyone explain how a company like Sequoia invests in something like FTX but never sees audited financials from, you know, a real auditor?

I've done a couple of VC funded companies, and if we had done anything one percent as egregious as SBF and his friends, we'd all be in jail.

You thought you bought software – all you bought was a lie

PghMike

Re: Perfect fidelity

Right you are. If I save and reopen an MS doc, and then try to add a new line to an existing drawing, it will *never* possible to draw a perfectly vertical or horizontal line again.

Try to change the formatting of a numbered list? It'll be close to a miracle if you pull it off at all. If it also looks the same next time you open the document, it's time to get the Vatican's Miracle Certifying Team to fly over.

PghMike

No perfect compatibility within Office either

For whatever reason -- operator error or terrible software -- I've never been able to insert a drawing in a MS document (word or powerpoint), save it, reopen it later, and then change it again perfectly. If I try to draw a new line, the grid has been recalibrated so that I can't draw lines parallel to those already in the drawing. Sometimes, even the old lines in the drawing move around a bit.

It's also disappointing how hard it is to draw even the simplest diagrams. You'd think after this many decades, it would require fewer clicks / keystrokes to insert a few text boxes and connect them.

Get over it: Microsoft is a Linux and open source company these days

PghMike

Mostly agree

I'm actually a MSFT employee, so dismiss what I say on that basis if you like. But all my work is on Ubuntu running in Azure, which I get to via a MSFT provided MacBook Pro. I haven't used anything running Windows (except as a hypervisor host) since I got to MSFT 4 years ago.

MSFT really is a pretty decent open source contributor. Yes, they still have proprietary stuff, but the "embrace, extend, extinguish" policy is long gone. When we use an OSS product, AFAICT MSFT contributes any changes back to the community in any case where that makes sense for the community.

Back in the '90s, I worked in a networking company that was continually screwed over by MSFT arrogance ("we're changing the network mini port driver interface yet again for no good reason, so FU"), so I get where you're coming from. And I don't deal much with the Windows folks, but the parts of MSFT I see try to make customers happy, not jerk them around.

RIP: Creators of the GIF and TRS-80

PghMike

nit on inflation

$599 in 1977 is closer to $2800 in today's dollars. This article *was* written in 2022, right? It's been 45 years.

The dark equation of harm versus good means blockchain’s had its day

PghMike

How to stop blockchain currencies

I'm pretty sure you can make it illegal to buy/sell cryptocurrencies, at least for banks. If you make it hard to convert money people actually have into crypto, you can reduce the amount of fraud enabled by crypto.

European silicon output shrinking, metal smelters closing as electricity prices quadruple, trade body warns

PghMike

Re: But...

I remember as a 9 year old at the '65 New York Worlds Fair, going to the GE pavilion to see the mock up nuclear fusion reactor (a few strobes under a glass bubble). I'm more surprised by how old I am than in the progress in fusion since then.

PghMike

turning off nukes

Maybe not the best idea to shutdown nuclear power plants early? I'm looking at your Germany :-)

A lightbulb moment comes too late to save a mainframe engineer's blushes

PghMike

Thanks for a great story!

Hacking the computer with wirewraps and soldering irons: Just fix the issues as they come up, right?

PghMike

Re: MagicSix?

Any sufficiently advanced technology is indistinguishable from Magic.

PghMike

Re: I'd hire him...

The term 'noise' is a bit misleading -- the main issue was that the data returned by an instruction that took a fault was *undefined*. The hardware wasn't unreliable, the OS designers just assumed it provided functionality (restartable 'segment missing' faults) that Interdata didn't actually provide until 3 years later.

They did eventually fix the issue, so that by 1978, the next model, the 8/32 worked as desired. But that was three years later, and we wanted a solution to share these machines much faster.

PghMike

Re: MagicSix?

Yup, MagicSix it is.

Fix five days of server failure with this one weird trick

PghMike

When I was a grad student, 40 years back, I remember the folks who built the research machines told me "It's always the power supply. Even after you've ruled it out, it's *still* the power supply."

That being said, the nightmare tech I remember is coax, specifically tapping receivers into a coax cable, before there was point to point ethernet cabling (this was 10 MBit ethernet, so we're going back to around 1980). You'd add a new machine to the network by essentially drilling a hole into the cable, and adding a tap that reached the copper core of the cable. Of course, each time you added a new machine, you'd get different reflections from all the previous taps, and of course you also had to terminate the cable itself properly. The end result is that you'd add a new machine, and some random arbitrary pair of other machines might start having trouble communicating.

Windows 11 comes bearing THAAS, Trojan Horse as a service

PghMike

Re: Forgive me for saying this...

Teams does work for simple meetings.

What drives me nuts is the maze of twisty passages that makes up your life in Teams. Someone shared a document with you during a chat? Good luck finding it ever again, even if you were smart enough to remember to pin the conversation. Even if you remember who it was who shared it, you have to find the specific conversation. And then you have to remember if they uploaded it to the 'Files' tab or if it was part of the conversation inline.

And sometimes the conversation is in the Teams tab, and sometimes in the Chat tab.

And random functionality requires you to click "Open in SharePoint," where you're in a similar 'app' but different. And of course, you have to guess that the functionality you're looking for is in sharepoint in the first place.

And again, search just plain doesn't find anything.

It's just a mess. And the way that MS works, shoveling functionality into applications with no thought of the user experience, it's only going to get worse.

I no longer have a burning hatred for Jewish people, says Googler now suddenly no longer at Google

PghMike

Re: I am lucky

There was a recently letter to Slate's "Care and Feeding" column where a letter writer wrote in that she recently noticed that someone who bullied her in a racist manner as a very young child was now, 30 years later, an elementary school teacher.

She wants to contact the school to report the teacher's behavior as, I'm estimating, a third grader.

That's insane, and I was shocked to see that the column's author thought reporting the teacher's behavior back when she was about 8 years old, was a good idea, but suggested going to the principal of the school instead of the school board.

And about half the comments to the column are of the "yes, report her," category.

PghMike

Too strong a response by Google

It's disappointing that it took so many years for this guy to get over his anti-semitism, especially since he's spent so much time in the US apparently still under the sway of his upbringing in Egypt.

But still, it is disappointing that Google would dismiss someone for testifying how they got over the hate they were raised with. Many of us grew up in way less tolerant times, and it is generally a good thing to hear about someone who overcame the biases they were raised with.

And yes, I'm Jewish, so don't call me an anti-semite for dissing Google here.

Google's diversity strat lead who said Jews have 'insatiable appetite for war' is no longer diversity strat lead

PghMike

Re: Out of context

His point seemed to be that emotionally he found homosexuality repugnant, but he recognized that logically his position was nonsense or worse. And as for his emotions, he wrote "Indeed, it is through the lasting nature of that friendship that my emotional core is changing."

IOW, in my reading, he knows his emotional reaction is bad, and is working to change it.

Blockchain may be the machinery of mischief, but it can't help telling the truth

PghMike

Why blockchain?

Couldn't Christies, and any other trusted name in the art world, just sign NFTs with a secret key, which could be verified by anyone with their public key? You'd save the energy burned by block chain updates, and still get certified authenticity.

And isn't the idea of NFTs for physical objects sorta silly? Aren't you signing a checksum of the object? And how do you generate a replicable checksum of a physical object?

This is all a bad joke.

We can't avoid it any longer. Here's a story about the NFT mania... aka someone bought a JPEG for $69m in Ether

PghMike

Dumber than a rock?

AFAICT, someone paid $69M to get a digital signature of a JPG added to a blockchain, indicating that they own the JPG. And it isn't even clear whether the new owner owns the copyright to the image, or is just the registered owner of a file that can be still be copied and distributed to others. Or even trivially modified by its creator and sold again.

Wow.

Microsoft says it found 1,000-plus developers' fingerprints on the SolarWinds attack

PghMike

Re: Oh those Russians!

Really, wake up. MSFT of 2020 is not MSFT of 1995. There's a huge amount of support for Open Source in Azure, and throughout MSFT in general. The MSFT IT department even supports MacBooks (I'm typing this on my MSFT provided MacBook, and no, it isn't running Windows) and iPhones.

Disclaimer -- I've worked there since 2018.

Web prank horror: Man shot dead while pretending to rob someone at knife-point for a YouTube video

PghMike

Re: This is why they should be banned.

Well, you can call it a prank, but the idiots did commit assault with knives (using the legal definition of assault: "putting another person in reasonable apprehension of an imminent harmful or offensive contact"). The difference between a prank and a robbery was unobservable to the victims, which is why one of the pranksters is dead.

Soon, no more blood tests or probing for prostate cancer? AI claims 99% success rate using more relaxing methods

PghMike

Re: AI?

It sounds like the weights of the various biomarkers is determined via machine learning, i.e. feed lots of examples through a neural network. This is typically called AI these days.

PghMike

Something's missing from the stats here.

There's something weird:

"Lee said that although the false positive and negative rates of the algorithms were low at 0.024 and 0.037 respectively, the team needed to verify their results with many more patients."

With 23 people tested, why wouldn't all of your error rates be multiples of .043 (1/23)?

Pop quiz: You've got a roomful of electrical equipment. How do you put out a fire?

PghMike

Nice title

Just another Pink Floyd fan....

Proposed US fix for Boeing 737 Max software woes does not address Ethiopian crash scenario, UK pilot union warns

PghMike

Re: Fundamentally flawed

I think you're right. The problems with this plane stem from the fact that the engines are too far forward, in front of the center of gravity. They did this because the fuel efficient engines Boeing wanted to use were too big to fit further back. Unfortunately, if the nose rises, wind hitting the front of the engine pushes the nose even higher, an example of positive feedback. Other planes don't have this problem, and so don't require an MCAS to be stable when the nose drifts upwards a bit.

Fundamentally, this is a flawed design, and the MCAS is a complex bandaid designed to hide this fact. The only solution is to put smaller engines on the plane, and live with the worse fuel economy.

As the pilots quoted here noted, once the MCAS fires and the plane dives, the pilots probably aren't strong enough to adjust the trim against the wind. It's pretty scary watching the FAA and Boeing make the same potentially fatal mistakes again, this time in public. Let's hope the BAA, CAA or European equivalent put a stop to this crap.

IBM ordered to pay £22k to whistleblower and told by judges: Teach your managers what discrimination means

PghMike

Re: £22K? Is that all?

If she's smart, she'll leave on her own soon. There are better companies to work for than IBM, and I'd bet that the local IBMers will be vengeful AF.

Start Me Up: 25 years ago this week, Windows 95 launched and, for a brief moment, Microsoft was almost cool

PghMike

Good summary of MSFT

Great summary, not only of Windows, but of MS in its entirety. Their approach was, until Azure, to just throw a random collection of poorly implemented functionality into Windows, all bundled together for whatever they charged (around $100 for the weakest tea, IIRC), cutting into the market for anyone who wanted to do a better version of whatever Windows threw in for free. I believe Ballmer called it 'cutting off the air supply' of their competitors, in an accurate turn of phrase that I believe he probably regretted.

Windows NT was the first system whose internals you could study without becoming ill, and that only if you didn't look at its VM or file system interfaces. I haven't dared look at it for years, but I have no reason to believe it's improved. SunOS did the file system / VM system much better than anyone else at the time (thanks, Steve).

MS definitely violated all sorts of anti-trust laws, but what really did them in was the Internet. They just didn't get it, and their browser's attempts at doing things proprietarily continued to hurt them, as their attempts to innovate were labeled, not incorrectly, as 'extend and extinguish'. It took Amazon's success with AWS to define the market well enough for MSFT to actually start competing again, with Azure, in an environment where, for the first time, they weren't leveraging their position in Windows to get an unfair advantage for mediocre technology.

PghMike

IIRC, and I'm not sure I do, I think the Windows 16 bit 'VM' might have been the only preemptively multitasked component in Win95.

Page: