Posts by EricM

NASA's X-59 plane is aiming for a sonic thump, not a boom

I must admit, beyond basic research I do not fully get the point of the X-59 program

Size:

This is a demonstrator, comparable to F16/F18, a 1-Pilot aircraft.

If you scale this design up to a 100 seat airliner (comparable to the Concorde), a lot of probably non-linear effects will determine the energy of the resulting sonic boom - which then might or might not be Concorde-sized.

Sound:

Even if they achieve a 30-50% reduction in energy carried by the shock wave, it's still a shock wave. Looking at FR24, continental US, Europe and parts of Asia have already such a dense coverage of overflights ( which you mostly do not hear at all today) that even reduced shock waves every few minutes would probably be pretty hard to tolerate.

Energy consumption:

Sub-sonic air travel today is energy-intensive as hell already. Raising speeds to supersonic will raise the energy usage and CO2 emitted to fly the same route from A to B by a hefty margin.

Given the realities of air travel: Why?

Why reducing a 4 hr flight (take-off to touch-down) by 1 hr, if traveling to the airport, waiting for security screening, waiting for boarding, boarding itself, off-boarding, waiting for luggage, traveling from the airport ... already take the better part of a day anyway?

Microsoft preps big guns to shift Copilot software and PCs

Customers struggling to find use cases? Easy: Just give marching orders to more sales-drones.

Honestly: I fail to remember any other groundbreaking innovation in human history that had to be rammed down customer's throats that hard ...

Something probably really has to give in 2025 ... and I for one think it's the AI hype itself.

Australia passes law to keep under-16s off social media – good luck with that, mate

Re: a workable age verification system

Exactly, by promising "that citizens will be able to use social media without having to show ID" IMHO law makers promise the only technical alternative that has at least a remote chance of achieving what was legislated, will not be implemented.

Occam's Razor and all that, but being _that_ stupid?

I'd expect that "promise" to be dropped right after the next election...

Panasonic brings its founder back to life as an AI

Virtually reviving dead people to give advice - another dystopic feature has transitioned from the realm of science fiction into reality.

Note to world: Neuromancer was not meant as an instruction manual.

Yup, half of that thought-leader crap on LinkedIn is indeed AI scribbled

Along those lines ...

... the other half of these "insightful" posts on Linkedin does not generally have much more value than the AI generated crap.

Kill Oracle's 'JavaScript' trademark, Deno asks USPTO

For clarification: This is an old SUN Microsystems Trademark from 1995

So it originated in a distant past when we all still used marble, hammers and chisels to code - and when SUN just had created the Java Language and more or less correctly might have assumed all things with the name "Java" in it (besides the island) to be their intellectual property. Oracle bought the mark together with the rest of SUN in 2010.

https://tsdr.uspto.gov/#caseNumber=75026640&caseType=SERIAL_NO&searchType=statusSearch

Section "Assignment Abstract Of Title Information"

2024, however, not contesting the cancellation of this inherited trademark would probably be the sensible thing to do for Oracle.

Imagine a land in which Big Tech can't send you down online rabbit holes or use algorithms to overcharge you

Re: The idea that "Echo chambers" are forbidden would kill some Social Media sites immediately

... luesky, Mastodon, Truth Social etc... Is that really a good thing, regardless of your political swing ?

If a "social"media site lives off the business case to create echo chambers that catch users to benefit their bottom line (or political leaning) and divide/hurt society in the process, then yes, I think it would be a good thing if they die. Regardless of their political affiliation.

This coming from the Chinese government that tries to run their whole domestic Internet as one big state-controlled echo chamber, is, hoever, beyond ironic.

The actual motivation there might be more protcting their domestic echo chamber from the effects of rivaling echo chambers.

So yes, I agree with "too Orwellian"

Europe glances Russia's way after Baltic Sea data cables severed

Re: What a damned stupid species we are

Agree on your post title, however, ...

> The really harmful ones are those that get into top political positions

No, the _really_ harmful ones allow themselves to be manipulated into voting psychopaths and sociopaths into power, while they still have a mostly functioning voting system.

Both Putin and Trump have in fact been elected in their first 2 elections. After Putin's 2nd term is was pretty much game over for any form of meaningful democracy in Russia.

Will be interesting to see if the US will be able to avoid a "King Trump"

Judges not impressed by Amazon, SpaceX's attempt to have NLRB declared unconstitutional

Oh, our favorite free speech proponent with so many fans and followers is again ...

...trying to suppress free speech that does not agree with him ...

Looks like he is firmly following the examples of Iran, Russia or N-Korea: You are allowed to freely say anything - if that agrees with your beloved leader.

The amount of cognitive dissonance required to actually be one of his fans seems to grow each day ...

Letting chatbots run robots ends as badly as you'd expect

Re: Where do you consider your organization in the AI maturity scale?

I miss the selection "Marketing in full overdrive, IT and engineeering desperately looking for real-world problems fitting to available AI 'solutions'"

Agree on different motivation

While Asimov probably has spent more time thinking about how to prevent damage to humanity (think: Foundation) than the current bunch of AI proponents combined, he also had no incentive to create actual damage - in contrast the current AI/Robotics industry.

Academic papers yanked after authors found to have used unlicensed software

The connection is copyright

Companies like Elsevier take tax-funded resaerch results, tack on their own copyright and make a living off selling access to that content.

So they have a conflict of interest with regard to copyright law, where they promote a very strict interpretation, basically to protect their own revenue stream.

In addition, not having "a license" to base a paper on can be anything from having no license at all, only having a personal license instead of a team license, up to not having bought additional rights to publish results/screenshots/logos/etc.

Clues to Windows Intelligence found in Windows 11 builds

Lack of consistent naming is IMHO the least confusing aspect of "AI" as part of an OS

Aspects of an LLM in notepad.exe that result in rephrasing suggestions that are often worse than their input and a Copilot nobody wants to spend real money on should have made it clear:

The LLM-in-everything variant of AI is currently a huge commercial bubble about to burst.

There is simply no real use case to have a more or less context-aware also-ran-grade text generation capability inside a general OS, because that capability is typically placed in front of the keyboard. At the same time all this AI-infused/infested stuff generates uncurable privacy nightmares just by the way it operates.

So what Microsoft is currently trying here, will be very good reasons to avoid Windows 11 at all cost as long as possible.

The billionaire behind Trump's 'unhackable' phone is on a mission to fight Tesla's FSD

10k lines of code?

Even if we go back to the simple times of a C64 or ZX81, 10k LOC won't buy you much functionality.

On a (modern) phone, initializing the hardware, radio, GPS, etc. will already take more than that.

So I'd like to see what definition of "OS" is being used here.

Is the actual functionality located then in "applications" that interact with the hardware, so in case of a security problem you can claim your OS is secure, because the problem was in one of the many applications needed to actually work with a device?

Overall I'd suspect this claim to be true only under a very specific set of circumstances.

Sysadmins rage over Apple’s ‘nightmarish’ SSL/TLS cert lifespan cuts plot

Re: OK, then let's focus on really strict security

> The problem they are trying to solve is certificate revocation.

Yes, they solve one of their problems by creating a host of problems for a lot of admins in all other organizations.

Applying full automation to cert changes is a) not possible in every HW/SW/compliance level environment and b) creates its own set of risks and attack vectors.

The CA-based "web of trust" is already broken to a certain extent. As multiple events of fraudulently issued certs showed, a valid CA-issued cert is no guearantee that you are really talking to the partner you think you are talking to. It never really was.

Every single of the hundreds of root CAs is a single point of failure in this web.

Throwing in full automation as a means of a "fix" will multiply the existing problems, not solve them.

OK, then let's focus on really strict security

Let's shorten the lifespan of certs to 48 hours, so stolen certs will basically expire immediately.

Or why not 12 hours? 2 certs a day should be worth it to stay secure, right?

Then, why not 5 minutes? Would be much more "secure" than 12 hours.

/s in case this wasn't obvious.

Somewhere along the lines I lost the plot on why shorter lifespans are regarded as more secure.

Sure, shortening 8 years to 3 seemed to make sense. 3 years to 1 year? Already questionalble.

But shorter? What's the actual meaning?

If a cert holder cannot be trusted to not spill their production certs for one single year, why trust them at all, you know, with all the rest of our business?

Europa Clipper heads to Jupiter: Can its icy moon support life?

Re: Fascinating

> Does anyone know why they can't just slingshot away from Jupiter and it's moons and send it off into the dark with it's sensors running sending NASA data until it runs out of power (a la Voyager)?

The Pioneers and Voyagers were fly-bys. They did not slow down to examine the system in depth, they kept their speed and went off to interplanetary space. No fuel needed for that.

In contrast, Europa Clipper will burn most of its remaining fuel at Jupter arrival to slow down enough to enter into an orbit around Jupter.

After that burn it sits deep inside the gravity well of Jupiter and its remaining fuel only allows slight trajectory corrections.

So there is not enough fuel left to accelearate enough to leave the Jovian system again.

Microsoft sprinkles AI 'magic' and additional storage tiers on OneDrive

Where is the legal Copilot use case worth actual money?

If nothing else, Microsoft's use of the term "magic" in this release should be understood as 747-sized red flag ...

I perfectly understand the use of AI based tools, usually not LLM-based, in science for example for advanced pattern matching/localization or variant analysis in very complex datasets.

However, in the commercial space mentioned in this article, the potential routes of applying current AI/LLM capabilities in a way that make serious customers want to pay for them seem to run dry.

So far we have seen some nice to have and some outright dangerous features branded "copilot".

They are used by most out of curiosity a few times, sometimes they work, sometimes they fail more or less subtily - and then, what?

LLMs greatest achievements so far are making a less capable individual fake average capabilities and helping scammers generate their sub-par content at larger scale.

So what gives?

What is the real ROI of that technology?

For the companies pumping billions into training and running the models (hardware, software, engineerring, energy )?

For the society as a whole that needs to learn to live with better equipped scammers and the increased energy consumption of AIs?

Or is this specific LLM-for-everything AI bubble finally about to burst?

Amazon CEO wants his staff back in the office full time

Sure, nothing drives the ability to move fast, clarify and invigorate a sense of ownership or decision-making like a forced relocation follwed by a 3-hour commute each day ...

Hint: If you really feel the urge to force especially the elder part of your workfore to resign, at least provide them the courtsy of inventing a somewhat intelligent sounding lie.

Meta back at it, harvesting Britons' public Facebook, Insta feeds for AI training

Re: Objection!

True, but not "being on FB" (I am not) does not even allow you to opt out of whatever data and meta-data they already have about you for example by scraping phone directories of Whatsapp users.

Admins wonder if the cloud was such a good idea after all

Short lived memories, maybe?

Maybe one effect might be that people tend to forget negative exeperiences from the past?

On-premise datacenters with limited capacity were resized in a matter of 10? years, if at all.

Often you were limited by available space, power, CAPEX and projects were sized to fit in whatever limited capacity was available.

Remeber the resulting amount of pretty expensive work spent on code optimizations in earlier days?

Now that all developers/admins have practically unlimited capacity at their fingertips, Optimization IMHO does seem to get much less attention nowadays.

So I find it unsurprising that cloud bills go up over time, however, lower manpower cost should balance rising cloud costs to a certain extent for many organizations.

In the long run, organizations need to find their balance between manpower cost for optimizing and cloud/on-premise capacity cost.

Musk's X, Media Matters headed to trial

Re: The other part of the story

That's why I prepended my statement with "formally".

Of course there are obvious indirect incentives for the judge to find in favor of X.

Re: The other part of the story

Formally, because Tesla - the company - is no party in this dispute.

Microsoft Bing Copilot accuses reporter of crimes he covered

Re: AI _is_ just overhyped statistics

OK, in case you understand the technology better, please feel free to describe it better here - in a way that clarifies the distinction between statistical weights and a real cause-effect understanding that LLMs are lacking.

AI _is_ just overhyped statistics

A string often occurs near to other strings? That's a signal.

A name often occurs near to other terms? That's a signal.

A name of a judge or journalist often occurs near to crimes? That's a signal.

AI models are - just slightly oversimplified - build on the rule: correlation _actually_is_ causation.

Seattle airport 'possible cyberattack' snarls travel yet again

RCA

> Is it more automation?

Well, more automation with poorly understood failure modes is exactly what brought us to the current status.

my personal Root Cause Analysis : The whole security thingy has been so fu**ed up in 3 simple steps:

1) Deploying knowingly insecure Windows OS to run critial systems - I'm in this industry since 1991 and I never understood how you can select any incarnation of Windows since then to run server workloads.

2) Trying to "fix Windows Security" with layers over layers of 3rd party AV/EDR snake oil that depends on high-frequency updates and introduces its own bugs and attack vectors, AI, Machine Learning, whatever to somehow remediate just the known Win security problems

3) Create general rules like PCI DSS, NIST, etc., that codify the resulting over-complex mess and make it mandatory for everyone, even those using sane operating systems. Note this is usually named "compliance", not "security", for obvious reasons ...

Result: A wrong update of a major EDR company in 2024 can take out a good part of the commercial Internet servers over night.

Not only Windows, Linux, too.

So we tried to minimize Risk A (Windows Security) by creating Risk B (DOS by EDR), which is more probabable, more severe and escalates easily to an international, industry-wide scale.

And it does not even really fix the original problem - Windows security.

Rethinking this might be a very good approach.

Rocket Factory Augsburg engine test ends in explosion at SaxaVord spaceport

Looks like a fuel leak above the engines?

That horizontal flame to the left starting directly after ignition does not seem to correspond to any of the engines.

Ruptured fuel line?

Gartner mages: Payback from office AI expected in around two years

Can we agree on a date when Paek Gartner happened?

I'll start: 2015

Since then they are sliding down the slope of obviousness. No enlightenment in sight ...

AI or bust? Only one part of US tech economy keeps growing, says analyst

I'd challenge that

> with a small number of very large companies that are exposed to cloud and AI doing much better than the industry average.

Depends on what you understand as "being exposed" to AI ...

Companies currently throw out "comes with AI" as a pretty empty marketing term, as most of their products do not in fact contain any working functionality based on LLMs, Neural Networks, Deep Learning, e.g. "AI" in the stricter sense of the word.

People are - sadly - buying into this stuff anyway, usually shortly before they discover they got sold some crap solution from yesteryear - just with a shiny, new marketing slogan that does not mean anything real.

So companies "exposed" to AI come in two flavors:

- A) Companies using "AI" as a marketing term: Those will be punished the next time the customer purchases something

- B) Companies that have fallen for "AI"-marketing: They are on a steep leraning curve to avoid companies using the term.

Group A is in fact currently doing better - by basically lying to their customers. So "doing better" now might well create a big liability which will hurt their future financial results.

Group B will be doing better in the future as elevated cynism will help to provide some hardening against future marketing BS.

Before we put half a million broadband satellites in orbit, anyone want to consider environmental effects?

Plus the effects of thousands of reentries burning up metal in the stratosphere and above

A re-entry so far was an occasional event for earth's atmosphere.

Every mega-constellation having about 5000 re-entries per year will increase the number of those events by 3 orders of magnitude.

Likewise the mass of burned up metals will increase, with as of yet unknown effects.

Not sure what the effect will be, but if there is one and we discover it only, when the constellations are all up, there will be nothing left we can do ...

Elon Musk claims live Trump interview on X derailed by DDoS

It's understandable to _pretend_ it's a DDoS

Sounds better than admitting you fired too many employees with a clue ...

Here we go again with more AI crime prediction for policing

> while AI systems can suffer from inaccuracy, high cost, bias, and other limitations, "there is consensus by the majority of the researchers on the importance of predictive algorithms on the policing landscape".

So: We know, it's expensive and inherently broken, but it's important?

What kind of BS argument is that?

I've heard politicians putting out this kind of nonsense, but "resarchers"?

Come on ...

IBM, Accenture, and Oracle, asked to develop Prometea, estimated the project would cost $100,000.

Surely there are some zeros missing...

For that kind of money those companies cannot even develop a cost estimate ...

Boeing's Q2 nosedive buoyed by appointment of new CEO

Boeing missed the mark on pretty much every analyst expectation in Q2

If Mr Ortberg does what needs to be done, e.g. fixing the way Boeing develops and builds planes, Boeing will miss market expectations for many years to come, as those are based on past performance, which in turn was based on ignoring gaping security problems in its development and build processes and a workforce who was trained to and rewarded for "looking the other way".

In other words: Boeing not missing the mark constantly for the years to come means that its development and production processes are not being fixed, but deficiencies continue to be concealed.

Zuck dreams of personalized AI assistants for all – just like email

Re: personalized AI assistants for all

"like mail" ... got me thinking.

Maybe a personal AI assitant reading all my corporate mails and summarizing them as "just the usual BS today" or attending 4 hrs of listen-only Teams calls while creating a summary of the interesting 30 seconds would be a _real_ productivity gain after all :)

Kamala Harris's $7M support from LinkedIn founder comes with a request: Fire Lina Khan

Microsoft board member requests FTC chair being removed - pays money to upcoming president

This approach to bribery seems to lack even the slightest trace amounts of subtility ...

Shuttle Columbia's near-miss: Why we should always expect the unexpected in space

Note to self: Schedule another review of my servers contingency plans...

'Gay furry hackers' say they've disbanded after raiding Project 2025's Heritage Foundation

Re: "Christian values" - Right Wing Evangelical flavour.

> yes, agreed. Believers in the various marxist cults are unconcerned with murdering millions. Pol Pot, Mao, Stalin.

Agree so far. Just as the border between evangelical and far right is thin, also some other political ideologies border close to irratioal belief systems.

> Their fellow totalitarians like

> inhuman riots activists,

> lawfare practitioners

> also white ant representative governments.

Not sure I follow the plot of your post - what are those groups? Who do you refer to exactly? Maybe try to explain with less slang.

>I wont mention the common epithet applied to those who also had same beliefs but oddly, are damned.

Complete lost plot here - what does that mean?

"Christian values" - Right Wing Evangelical flavour.

Organizations like the Heritage Foundation show pretty drasticly why religion in general is a dangerous invention.

Everyone and every organization can claim to speak in the name of the imaginary being accepted by a majority of a group as a god.

It obviously even works if this person or organization claims statements or policies that are completely contrary to any guidance the being of choice has provided in the past - otherwise Right Wing Evangelicals would not exist.

Religion in general is like a gun lying fully loaded on the street for everyone to pick up and start shooting.

There is virtually no overlap between the spirit of the new Testament and policies of right-wing Evangelicals, but believers seem to be able to tolerate extremely high levels of cognitive dissonance if someone convinces them, their "god" wants it.

"Deus vult" still works far too good in the 21st century.

BT bets big on AI with ServiceNow to cut legacy baggage

Translation:

BT Group expected to discover the usual differences between GenAI technology marketing and reality...

Chinese space company accidentally launches rocket in test gone wrong

Let me suggest unscheduled ballistic disassembly for this failure mode

Nvidia loses a cool $500B as market questions AI boom

market questions AI boom - about time this hype gets dialed down to normal levels

I'd propose a vaporware award for genAI for constanly delivering wave after wave of huge promises/demos/betas while delivering not one _really_ usable product that works reliable and repeatable.

Mozilla is trying to push me out because I have cancer, CPO says in bombshell lawsuit

In what direction is Mozilla heading?

https://www.theregister.com/2024/06/21/firefox_127_private_window/

https://www.theregister.com/2024/06/18/mozilla_buys_anonym_betting_privacy/

Just in the Last week they bought an ad company to "bring meaningful adverts to firefox" - f*cked up the privacy setting and now this ...

I must say, based on my end-user viewpoint, the former technical and strategic leads (CPO among them) seem to have done a much better job than the current ones.

Independent from that:

If only half of the allegations with regard to handling the illness of a staff member in this article are true, Mozilla should probably fire their CEO instead of the CPO.

Privacy features lose their way in latest Firefox update

Re: "Weapon of choice" nails it pretty well

In case that happens, I'll probably treat it the same as "optimzed for Internet Explorer" at the time: I don't give a shit about those websites.

"Weapon of choice" nails it pretty well

NoScript without limitions imposed by Chromium ..

Qilin cyber scum leak data they claim belongs to London hospitals’ pathology provider

Re: Infrastructure should be more resilient than this.

> Air gap it

How do you think healthcare providers would communicate?

Pen, Paper and Mail? The rate of transmission errors and time-outs in such a system would probably be the #1 reason of death of patients.

Operating these systems connected to the Internet is serving a purpose.

> Infrastructure should be more resilient than this

That one I agree to.

However, implementing security concepts (or any concepts that provide more resilince) costs money.

And what company spends additional money on security when the provider of the brand-new software you just purchased, told them it will be "secure"?

And what provider doesn't do that?

Meta warns bit flips, other hardware faults cause AI errors

Re: I'm a bit out of touch with the hardware design

My first thought, too, but the effects of ECC would be a bit too obvious to omit is such a statement, wouldn't it?

Are they maybe talking about bit flips in the CPU itself, in registers, logic units, controllers, etc?

But for those the error probability they come up with would seem to be astronomical ...

We basically would not even have non-AI universal computing with this error rate.

Oracle Java license teams set to begin targeting Oracle users who don't think they use Oracle

A "special" kind of relationship ...

"They don't have a relationship with Oracle. But Oracle has tracked Java SE downloads to their company. "

Or so they might at least say...

Make sure you _really_ do not use Oracle Java - there is literally not a single reason not to run whatever Java stuff is needed on OpenJDK these days, which needs exactly zero licenses from Oracle - and then tell Oracle to get lost.

This is not to say that Oracle might not be trying to muddy the water by talking just about "Java" instead of their outragoulsly priced version of "Oracle-infested Java (TM)"

Stanford Internet Observatory wilts under legal pressure during election year

No one does like being caugh lying.

I find the whole "free speech" angle to this fascinating.

In older times "free speech" meant you are able to express your opinion without fear of beeing punished for it.

However you were also free to bear the consequences of your speech, e.g. others discussing your staments and call you an idiot or liar.

Today, the "free speech" argument is more and more used to supress discussion about statements one has made and to avoid the consequences.

Exposing wrong facts, dissenting opinions, even calling a liar a liar is free speech as well.

So in fact "free speech" in this case is used as argument to effectively limit free speech, not only of the political opponent, but also of neutral observers.

However, the fact that the US conservatives are using this tactic is quite understandable.

There simply is no longer any rational argument they could use to ask voters to vote for their broken agenda or their idiotic candidates.

Since their only tactic left is lying to the general public, suppressing dissent and neutral observers becomes essential...