Posts by Decay

Anthropic's super bug-hunting model Mythos is hardcore good at math, as latest vuln under attack shows

I think people are getting rather hung up on angels dancing on the head of a pin here.

Whether Mythos genuinely derived the exploit from first principles, remembered something from training data, found an old blog post, or reused existing exploit code is interesting if the question is “how good is Mythos at mathematics?”

But from a security perspective, I don't think that is the important question.

The important bit is that an LLM can read an enormous amount of source code, documentation, CVEs, papers, forum posts and prior exploit material, retain the useful bits, correlate things that may be scattered across completely different places, and get from “there is something odd here” to a usable attack chain far faster than a human researcher or even a team of humans reasonably could.

Of course the fact that Math.random() is not cryptographically secure was already known. So was xorshift128+. So were the techniques for reversing it. So was Z3. None of that is the interesting part.

The interesting part is joining those dots in the right order, in the right codebase, and doing it quickly enough to matter.

Humans work exactly the same way. Very few security researchers independently invent every mathematical technique, cryptographic weakness or exploitation method they use. They read papers, documentation, CVEs, GitHub, Stack Overflow, mailing lists and other people's work, then apply that accumulated knowledge to a new problem.

The difference is scale and speed.

Saying “it didn't discover anything, it just looked it up” rather misses the point. Yes, perhaps it did.

But it looked up, understood, correlated and applied the right things out of an absurdly large body of information, and produced a useful result.

That is the capability worth paying attention to.

Whether Mythos got to step four by independently reinventing the maths or by remembering somebody's code from three years ago is an interesting academic argument.

From a security point of view, I care rather more that it got to step nine before we did.

Excel Canvas offers to turn your workbook into the boss's next report

So, if I have this right:

We had Excel.

Then Excel got complicated enough that we needed Power BI to turn spreadsheets into dashboards.

Now Power BI has got complicated enough that Microsoft is putting dashboard generation back into Excel.

And Copilot will apparently decide which metrics matter, which chart you wanted, and why the x-axis looks completely insane.

I can actually see the appeal. Give it a workbook, tell it to make something presentable, and out pops a report that refreshes as the data changes.

But there is an obvious question here.

Wasn't that what Power BI was supposed to do?

I assume the final architecture is:

Excel has the numbers.

Power BI has the model.

Fabric has the data lake.

Copilot has an opinion about all of it.

Canvas turns it into a dashboard.

Then someone exports it back to Excel so they can actually use it.

Bloody turtles all the way down!

Starship finally makes orbit but engine trouble cuts the victory lap short

I think that wording makes it sound more confusing than it actually was.

The original plan was deliberately to finish the main ascent burn on a passively safe suborbital trajectory. In other words, if they did nothing else, or subsequently lost control of the vehicle, it would naturally re-enter over a predetermined uninhabited area rather than remain in orbit.

They would then check that everything required to get back out of orbit was healthy and, only if they were happy, perform the small additional insertion burn to enter a stable orbit.

When one of the Raptors shut down early, the remaining five burned longer to compensate. That wasn't to ensure Starship would return to Earth. It was to get it to the velocity and trajectory they had originally intended at engine cutoff. If anything, failing to make up the lost velocity would have caused it to come down sooner, not remain in space.

The interesting decision came after that. They now had an engine failure and had to decide whether it was an isolated engine problem or evidence of some wider propulsion, propellant or vehicle issue. Once they committed to orbit, passive re-entry was no longer the safety net. They had to retain the ability to perform the de-orbit burn later.

Apparently the flight team were satisfied that the failed Raptor was an isolated issue and that they had sufficient redundancy to continue, so they went ahead with orbital insertion. That also allowed them to tick off two pretty major milestones: Starship reaching orbit for the first time and deployment of the V3 Starlink satellites.

The original plan was then roughly six orbits over nearly ten hours, which would also have provided useful time in the actual orbital environment, including repeated thermal cycling between sunlight and shadow.

Instead, after deploying the satellites they brought Starship home much earlier, performing the de-orbit burn after a little over two hours rather than completing the planned six-orbit mission.

So the interesting unanswered question isn't really why they went into orbit after losing an engine. SpaceX had specifically designed a decision point for exactly that situation. It's why, having decided the vehicle was healthy enough to enter orbit and deploy the payload, they subsequently decided not to leave it there for the full planned mission.

Could simply have been an abundance of caution after the earlier engine issue, or they may have seen something else they didn't like. Until SpaceX publishes more detail, we are all in the dark.

BOFH: How our Covid ransomware protocol's Y2K blockchain lowered uptime

Re: missing … the onions in our belts

To impart a malodorous aroma to the meeting that results in a quick evacuation of the room?

The only difference between the PFY report and a fair number of earnings reports is the scale of audacity and salaries involved.

SpaceX just about nails Starship test flight 13

That is fair, and I do not dispute that Musk himself is a genuine risk to the programme. His timelines are largely works of fiction, his management style appears brutal and erratic, and SpaceX has benefited enormously from government support and government contracts.

He may be a modern-day Henry Ford or Howard Hughes, take your pick: a transformative industrialist who is also personally objectionable, dreadful to work for and quite capable of becoming a liability to his own companies.

What irks me is the evident pleasure some people take in every failure, as though a rocket exploding proves not only that Musk is a charlatan, but that the people working for him are incompetent and that no progress has been made.

That is a peculiar way to judge an experimental development programme. A failed test may expose a serious design problem, but it can also prove that several earlier problems have been solved and allow the programme to move on to the next one. Failure and progress are not opposites in this kind of engineering. Quite often, the failure is how the progress is made.

Where I also disagree is with making a working commercial service the test of whether the rate of progress has been remarkable. Those are two different judgements.

Starship may yet fail commercially. It may prove too complicated, too expensive or too dependent on government and Starlink money. Orbital refuelling, rapid reuse, reliable payload delivery and upper-stage recovery remain unproven. None of that changes what the vehicle can demonstrably do today compared with what it could do three years ago.

Government support does not erase that progress either. NASA was plainly essential to SpaceX’s development, but there is a distinction between public support, milestone-based development funding and paying a company to deliver an actual service. SpaceX eventually converted that support into working Falcon and Dragon systems carrying cargo and crews. Starship has not crossed that line yet, but SpaceX has at least demonstrated that it is capable of doing so.

“We have seen this before” also cuts both ways. We have seen Musk promise self-driving cars, solar roofs and Hyperloop revolutions that never arrived on anything resembling his timetable. We have also seen SpaceX take booster recovery from an implausible ambition to routine operation.

My objection is not to scepticism about Musk or Starship’s commercial prospects. Both are entirely reasonable. It is to the almost gleeful assumption that every setback vindicates a pre-existing view of Musk and, by extension, dismisses the competence and achievements of thousands of people working beneath him.

Judge commercial success when they have a commercial service. Judge Musk’s predictions with all the cynicism they deserve. But judge the engineering progress by what the machine can do now compared with what it could do before.

An explosion may prove that the vehicle is not ready. It does not automatically prove that nothing has been achieved.

Yeah probably projecting.

I am honestly a little surprised to see so much of this attitude here. The readership of this site is largely made up of people either engaged in the trench warfare of IT or standing close enough to hear the swearing. Anyone who has spent time in those trenches knows how much planning, discovery, testing, rework and plain hard graft sits behind a successful result.

They also know how they would feel if, after months or years of work, some passing sage wandered in to ask, “I can’t believe you haven’t finished yet,” or, “What do you mean the milestone is amber?” Then, once the project finally succeeded, the same person waved away the entire effort as luck.

Most people here would be rightly incandescent if it were their project being dismissed like that. It is odd how readily some forget this when the engineers happen to be working on rockets rather than databases.

The comments put me in mind of a certain class of IT manager: the sort who is missing whenever a difficult decision is required, then turns up at the final project meeting to ask, “How hard can it be? Sure, it’s only an upgrade.”

The breed is not rare. It flourishes in air-conditioned meeting rooms where the coffee is free and every awkward action can be assigned to somebody who is not present. It feeds on status reports, breeds in steering committees and has an uncanny instinct for joining successful projects shortly before the photographs are taken.

Its natural counterpart is the armchair expert, who arrives after the difficult work has been done, explains that the answer was obvious all along, and takes the absence of disaster as proof that none of the planning was necessary.

In this world, complex engineering consists of pressing the large green **Upgrade** button, waiting for the cheerful bong and going for lunch. If no such button exists, it is probably because the engineers have hidden it to make themselves look busy.

Architecture, technical discovery, design, prototyping, testing, failure analysis, project management, risk management, vendor coordination and the actual business of building the thing are apparently just padding. Fussy little details invented by people who lack the clear strategic vision that comes from knowing bugger all about the subject.

There is no problem so complicated that it cannot be made simple by someone who has misunderstood it entirely.

A failed test is not necessarily evidence of incompetence. In experimental engineering, finding out precisely how and where something fails is often the point. You build it, test it, break it, inspect the remains, alter the design and have another go.

This is known as engineering. To the confidently uninformed it can look disappointingly unlike magic. Magic requires no project plan, produces no telemetry and has the great managerial advantage that nobody asks the wizard for a root-cause analysis.

A little historical perspective would not go amiss either.

NASA began developing Saturn V in 1962. Its first complete flight was in November 1967, nearly six years later, and Apollo 11 landed on the Moon in July 1969. That followed the Saturn I and Saturn IB programmes, years of ground testing and the expenditure of an enormous amount of money, effort and brainpower.

The Space Shuttle was approved in 1972 and did not fly to orbit until 1981. NASA and its contractors carried out more than 650 test firings of the main engines before STS-1.

Presumably nobody had yet told them it was only a reusable rocket. Had a sufficiently senior manager with no relevant knowledge been available, the entire business could have been wrapped up by Thursday afternoon after a workshop called **Simplifying Delivery**.

Starship’s first integrated flight in April 2023 reached about 39 kilometres, lost engines, tumbled and was destroyed. Eighteen months later, SpaceX caught a returning Super Heavy booster with the launch tower while the upper stage completed a controlled re-entry and splashdown.

Since then they have flown new versions of the booster, ship, engines and launch pad, reflown hardware, deployed test payloads and brought Starship back through re-entry in increasingly good condition. Flight 13 deployed 20 working Starlink V3 satellites and completed an intact upright splashdown.

The booster return did not go properly, because reality remains stubbornly unimpressed by the PowerPoint deck. Physics has attended every steering committee, accepted none of the actions and continued doing exactly as it pleases.

Starship is not finished. Routine orbital service, upper-stage recovery, orbital refuelling and rapid full reuse still have to be demonstrated. SpaceX also has the benefit of seventy years of accumulated rocket knowledge and its own experience with Falcon.

Even allowing for all that, the rate of progress is remarkable. In a little over three years they have gone from a full stack that could not complete staging to catching and reflighting boosters, deploying payloads and returning the largest spacecraft ever flown through near-orbital re-entry intact.

The armchair expert sees thirteen flights and counts explosions. The engineer sees thirteen increasingly ambitious tests, thirteen sets of telemetry and a machine that can now do things it plainly could not do three years ago.

One is looking at the evidence. The other has found the comment box, a marvellous little device that dispenses unlimited certainty without requiring the user to insert any knowledge first.

Criticise Musk all you like. There is no shortage of material. But dismissing the work of thousands of engineers as “luck” is the intellectual equivalent of arriving on the Monday after a two-year system migration, discovering that everyone can log in, and asking what took so long.

The great gift of the armchair expert is that no problem seems difficult once somebody else has solved it.

Good engineering makes the result look obvious. Eejits then confuse “obvious afterwards” with “easy all along” and generously explain this to the people who did the work, generally from somewhere well outside the blast radius.

SpaceX to try its luck again with Starship Flight 13 after engines and weather say no

Re: Public company woes

SpaceX has been quoted as saying that Starship V4 could eventually put 200 tonnes into orbit, but let’s leave that to one side and stick with 100 tonnes for now.

What could you actually do with a 100-tonne payload?

The military is obviously very interested. The ability to move 100 tonnes of equipment, supplies or munitions halfway around the world in a very short space of time would have military logistics people drooling.

That becomes even more interesting if you eventually have several launch sites and can launch every six hours, never mind every hour. There is a fair gap between saying that and actually doing it, of course, but you can understand why they are paying attention.

The bit that really interests me, though, is manufacturing and materials science.

Most orbital research today involves small samples and fairly compact experimental equipment. With a 100-tonne payload, you could start talking about something closer to an actual pilot plant:

* furnaces, vacuum chambers and crystal-growth equipment;

* robotic handling and quality-control systems;

* serious electrical and thermal-management equipment;

* tonnes of feedstock;

* spare parts and redundant machinery;

* shielding and containment for anything particularly unpleasant.

Microgravity changes how materials behave. You do not get the same buoyancy, sedimentation or gravity-driven convection that you do on Earth, which can mean more uniform mixing, solidification and crystal growth.

NASA and others are already looking at this for alloys, semiconductors, optical materials, pharmaceuticals and various other advanced products.

At the moment, a lot of that work comes down to producing tiny samples and asking whether anything useful happened.

A 100-tonne launch changes the question.

Instead of asking:

“Can we make a few grams of something interesting?”

you can start asking:

“Can we build a proper pilot line and make enough of it to find out whether there is an actual industry here?”

You could also build research facilities that are simply too large for the ISS:

* variable-gravity centrifuges;

* large electromagnetic levitation furnaces;

* long-duration alloy-solidification experiments;

* high-vacuum deposition lines;

* containerless processing systems;

* biological tissue-production facilities;

* laboratories comparing microgravity, lunar gravity, Martian gravity and Earth gravity.

The value would not only be in producing better materials.

It would give us a far better understanding of how gravity affects crystallisation, fluid flow, phase separation, combustion and solidification. Even where the final product is still made on Earth, the knowledge could be extremely valuable.

There is, however, one fairly obvious problem.

The current Pez dispenser is a Starlink delivery system. It works because SpaceX controls both the rocket and the payload and can design the satellites around the slot.

It is grand for launching stacks of Starlinks.

It is not much use if your payload is a factory, a telescope, a large pressure vessel or some awkward great lump of machinery.

SpaceX has previously shown a general-cargo version with a large clamshell opening and a tilting payload adapter. That would allow a large payload to be released rather than posted out through a letterbox one piece at a time.

For now, though, that remains a concept rather than an operational system.

So Starship may be approaching the ability to lift 100 tonnes, but it does not yet offer anything resembling a general-purpose 100-tonne freight service.

The other fascinating possibility is launching the factory rather than the finished structure.

A large payload could carry machinery and feedstock capable of producing:

* very long beams and trusses;

* enormous radio antennas;

* optical telescope structures;

* pressure vessels and propellant tanks;

* kilometre-scale solar arrays;

* rotating habitats;

* sunshades and radiation shielding.

None of those structures would ever need to support their own weight on Earth.

They could be extremely light, thin and delicate and still work perfectly well in orbit.

Instead of building some enormous structure on Earth, folding it into a rocket, shaking the life out of it during launch and then hoping it unfolds properly at the other end, you launch the factory and rolls of material.

The factory builds it in orbit.

That, to me, is where this starts becoming genuinely transformative.

It is not simply a case of putting 100 tonnes into space.

It is the possibility of putting the beginnings of an industrial base into space.

And if SpaceX gets anywhere close to the cost per kilogram it is aiming for, that opens up possibilities in materials science, manufacturing and engineering that we have barely started to think about.

These exciting times.

Health board apologizes for phishing staff with with bogus vacation day

Re: I thinking 'phishing' training is out of date

While I don't disagree with you, equally your CFO asking for Apple gift cards to complete the transaction is usually a give away. But you are right, a sophisticated scammer leveraging a compromised inbox even if it is a vendor or clients can salt an email with contextually correct speaking points that align with the victims knowledge of who and what is going on. Add in a text message asking the victim to read their urgent email and/or a voicemail in the CEO's voice and it becomes very persuasive. And a pain in the rear end.

You now have to read the email, judge it's veracity, potentially call, email or IM the sender by another means to confirm the email before acting on it. And sure the Apple gift card one is obvious but the click here to view the outstanding invoice email? I have told several vendors to stop sending me links to invoices, docusigns or other similar docs unless we have been discussing same previously. Worst offenders are companies that either outsource their accounting function or use a different domain name for different functions in the company.

And sure sitting in the comfort of your own home it all sounds like not much effort, but people are busy, running from meeting to meeting, doing work and in the 5 nanoseconds of free time see an email from a supposed trusted source, it looks at first glance to be legit, right sender, right context, damm did I forget to approve payment of that invoice? Let me click on the link to see what its for and bamm! you have failed at the first hurdle.

Want to really get those click numbers up? Send a phishing email to a company when it's bonus time with a description of why their bonus has been cancelled due to (insert likely reason here) and watch the CxO team click, authorize the MFA and grant access immediately. :(

Execs are the worst offenders, as a group they fail phishing tests more than anyone else because Masters of the Universe.

And we haven't got to real professionals who do their research, know your kids name, the school they go to etc. to hit you up with a really good/bad urgent call to action.

Amazon owns up to using 2.5bn gallons of H₂O in its bit barns last year

One nuance on the “just use the water directly, warm it slightly, and return it” argument is that in a lot of places that water is no longer treated as simple river water once it has gone through an industrial process. Even if it never touches the process itself and is only used through a heat exchanger or water jacket, the discharge can still become a regulated industrial return flow.

I am aware of one case where a company had to heavily prepare the water before industrial use: filtration, reverse osmosis, the works. That treated water was then used to cool a chemical reactor through a water jacket. Think three-storey-tall reactor, not a kettle in the break room.

The water never touched the inside of the reactor, but before it could leave the site it still had to go through a whole process to be tested, monitored, documented, and effectively restored as close as required to its original condition before being returned to the watershed.

I get why the rules exist. A brief glance at DuPont and PFAS is enough to explain why nobody gets to say “trust us, it only touched the outside.” But from an engineering and operational perspective, it turns a simple-sounding idea into a full regulatory palaver very quickly.

Americans would rather have a nuclear plant in their backyard than a datacenter

When MS was building a new datacenter in Dublin in the late 2000's the commentary was hundreds of jobs. Yes there were, while it was being built. Once it went live not so much. 15 jobs by RTE reckoning https://www.rte.ie/news/business/2007/1106/95616-microsoft-business/

Later in 2016 an expansion of the site touted was 1800 jobs!!!! But once live 140 jobs. https://www.independent.ie/business/technology/microsoft-given-the-green-light-for-four-dublin-data-centres-in-900m-project/34710317.html

So any job numbers quoted for any datacenter should be parsed carefully for construction versus post construction jobs.

One real reason AI isn't delivering: Meatbags in manglement

The sad reality is most users, business departments and management want a no thought required solution to a problem and AI is sold as that solution. I want AI version x, y or z to do this work for me faster and cheaper without putting any effort into thinking about how it will do it, what we are doing that could be improved. Just set it up and we will push the button and it all just works. This mindset isn't limited to just AI. We have all been victims of sales drones convincing the powers that be that product X is plug and play, just install it and your accounting system, CRM etc. it will be fantastic, just sign this contract and expect invoices every month.

The difference with AI is that the hype, and the decision makers experience with AIs writing grade A management waffle emails and memos, has convinced many that if it can write my emails so well, then just imagine what it could do building reports, performing complex analysis etc. I have somewhat solved this problem by having same people perform some basic analysis, stats or reports using a fairly simple dataset that they understand. Once the AI has mangled it a few times, made obvious errors or just plain made stuff up they usually respond with shock or occasionally with it must be that particular AI, so we try a few other flavors and they soon realize how bad it can be. Doesn't stop the sales drones trotting out the "our product is trained specifically for this scenario" bull but at least it gets the users or management asking questions and doing some critical thinking which seems to be a lost art.

US Navy pledges $448 million to test if Palantir is seaworthy

Wasn't Prince2 project management made for just this?

Canadian data order risks blowing a hole in EU sovereignty

Re: Treaties

Or freezing peoples bank accounts because the government is feeling aggrieved

https://www.forbes.com/sites/siladityaray/2022/02/23/canada-begins-to-release-frozen-bank-accounts-of-freedom-convoy-protestors/

"Isabelle Jacques, Canada’s assistant deputy minister of finance, told lawmakers on Tuesday that a vast majority of the locked accounts are now in the process of being released.

Jacques told a parliamentary committee that up to 210 bank accounts linked to the protestors—with cumulative holdings of C$7.8 million ($6.12 million)—had been frozen under the country’s emergencies act."

https://www.jurist.org/news/2024/01/canada-dispatch-federal-court-judge-rules-government-lacked-authority-to-invoke-emergencies-act-over-freedom-convoy-breached-charter-rights/

"Further, the court concluded government violated the Charter rights of convoy participants who had their bank accounts frozen. Mosley ruled that the federal government’s order empowering banks to freeze the accounts of those violating the emergency regulations and directing banks to provide violators’ information to the RCMP constituted an unreasonable search and seizure. The justice reasoned that the freezing of accounts constituted a “seizure” and that the order to provide the RCMP with blockade participants’ banking information was a “search.” Mosley held that the searches and seizures were unreasonable as they were not carried out according to an “objective standard.”".

70-hour work weeks no longer enough for Infosys founder, who praises China’s 996 culture

I don't know what all the fuss is about, it's an easy fix.

Step 1 chain your employees at the ankle to their workstation.

Step 2 Get a large muscly bloke at the top of the room with a big drum

Step 3 beat the drum to a regular cadence

Step 4 beat the drum faster when more "work" is required.

This looks fantastic on my spreadsheet where I can adjust inputs like food and water, performance vs BPM, solved the whole WFH nonsense my means of hardened steel shackles.

As a result my KPIs are all in the green, bonus time is looking good, any board related questions on productivity can be graphed and PowerPoint'd to death.

SpaceX loses debut V3 Super Heavy in ground test mishap

Re: Bad week for Elon

If true, and I have no reason to doubt it, it would be awesome to see both the company and the lawyers themselves sanctioned for this behavior. Too many corporate lawyers forget their oath and in particular their ethical responsibilities.

"I do solemnly swear: I will support the Constitution of the United States and the Constitution of the State

of Florida; I will maintain the respect due to courts of justice and judicial officers; I will not counsel or

maintain any suit or proceedings which shall appear to me to be unjust, nor any defense except such as I

believe to be honestly debatable under the law of the land; I will employ, for the purpose of maintaining

the causes confided in me such means only as are consistent with truth and honor, and will never seek to

mislead the judge or jury by any artifice or false statement of fact or law; I will maintain the confidence

and preserve inviolate the secrets of my clients, and will accept no compensation in connection with

their business except from them or with their knowledge and approval; To opposing parties and their

counsel, I pledge fairness, integrity, and civility, not only in court, but also in all written and oral

communications; I will abstain from all offensive personality and advance no fact prejudicial to the

honor or reputation of a party or witness, unless required by the justice of the cause with which I am

charged; I will never reject, from any consideration personal to myself, the cause of the defenseless or

oppressed, or delay anyone’s cause for lucre or malice. So help me God. "

And competition is a good thing. I think Blue Origin successfully landing the other day was a huge milestone. While it is a shame both companies are lead by twats who need a good slapping, the overall health of mankind's pace to exploring space are far better today than 15 years go, by an order of magnitude. This is history book stuff.

Last time I checked, you test stuff to see where it breaks. You do your best to make sure it’ll pass, but if you don’t push it, the unknown unknowns will bite you sooner or later. This one didn’t even trash the stand, and with the speed they’re turning out boosters now, it’s more of an annoying pothole than a real problem.

And let’s not pretend Falcon didn’t have plenty of bangs. Falcon 1 blew up three times before it made orbit. Falcon 9 lost CRS-7 halfway up, AMOS-6 blew itself to confetti during fueling, and a few Merlins have randomly let go mid-flight over the years. Early landings were basically fireballs with shrapnel. Every one of those screwups fed straight into the next revision, and that’s how Falcon ended up being the ridiculously reliable machine it is now.

Starship’s just doing the same thing, but scaled up. Break it, figure out what snapped, build the next one better. Same playbook, louder booms.

And the money situation couldn’t be more different. Back in the Falcon 1 days, each failure was basically 'one more of these and we’re dead.' After the third loss they were almost out of cash, and the fourth try was literally all-or-nothing. Now? They’ve got Starlink cashflow, NASA contracts, commercial launches queued up, and enough runway to eat a few prototypes without anyone panicking. What used to be existential is now just R&D overhead.

And yeah, credit where it’s due. The engineers and everyone on the ground doing the actual hard work deserve a nod. They’re the reason any of this stuff flies.

If you want a historical parallel, it’s closer to Henry Ford than the scrappy early-days Musk. Ford didn’t invent cars, he just figured out how to build the hell out of them. Musk is doing the rocket version of that, industrializing the whole thing. Falcon was the Model T moment; Starship is the assembly line moment. Ford had busted engines and broken prototypes everywhere too, he just didn’t have thousands of people watching on livestreams.

And let’s be honest: neither Henry Ford nor Elon Musk are exactly people you’d hold up as moral examples. Ford was openly antisemitic and ran his factories like a dictatorship. Musk… well, take your pick from the headlines. Both of them can be pretty awful as people, but they’re annoyingly good at building things and dragging industries forward whether anyone likes them or not. You don’t have to admire the man to acknowledge the engineering.

You'll never guess what the most common passwords are. Oh, wait, yes you will

I like the sites that review your password and refuse to accept it because it is too simple, wouldn't take much effort to add a database of top 500 common passwords and blacklist them. I get peeved when a site wont let my password manager use a 16 character password that contains uppercase lowercase numerical and special characters and then I have to tell the password manager to dumb it down.

Yes my password manager password is a lonnnggggg passphrase but its muscle memory now and I change it every so often.

Amazon brain drain finally sent AWS down the spout

Re: last dns outage I had

"What's with the thing against MBAs?

I see it all the time, yet there's never any explanation why this specific qualification is problematic."

As some others have said not all MBAs are bad, I would argue that experienced industry people with good knowledge and ability in their chosen field who move on to MBAs are usually quite good. These are rare unicorns.

Some light reading for you

https://evonomics.com/want-to-kill-your-economy-have-mba-programs/

https://www.advisorpedia.com/viewpoints/mba-thinking-can-ruin-businesses/

https://rpc.cfainstitute.org/research/cfa-digest/2014/10/you-reap-what-you-sow-how-mba-programs-undermine-ethics-digest-summary

https://hbr.org/2005/05/how-business-schools-lost-their-way

https://www.theguardian.com/news/2018/apr/27/bulldoze-the-business-school

Ok too much reading?

Toys "R" US

How management dropped the ball

Heavy debt burden: After a leveraged buy-out in 2005, the company carried large debt, which limited its flexibility.

Slow adaptation & innovation: TRU failed to keep up with the shift to online retail, changing consumer behaviour, and competition from big-box and e-commerce players.

Focus on cost/finance instead of investment: Because of the financial burden and perhaps a bias towards cost metrics, the company under-invested in store experience, technology, and e-commerce capabilities.

Strategic complacency: The business model (“toy supermarket”) became obsolete, yet the management did not pivot fast enough.

MBA fingerprints all over it

Short-term/financial driven: The debt-heavy structure and focus on servicing it limited long-term investment.

Metrics over context: TRU arguably prioritized profitability, cost structure and upkeep of the model, rather than deeply rethinking the changing retail/consumer context.

Tool/structure focus rather than adaptability: Despite being in a dynamic market, the company did not sufficiently embrace emergent strategies or culture change.

The takeaway?

Even strong brands can be undermined if management emphasizes financial engineering, cost metrics and static business models rather than continuous customer-/environment-led adaptation.

Boeing 787 Dreamliner

How management dropped the ball

Extensive outsourcing: Boeing outsourced a large portion of manufacturing and global supply chain tasks (approx. 70% of work) expecting cost savings and efficiency.

Complexity and control issues: The global network of suppliers resulted in loss of direct control, inadequate supply-chain visibility, quality issues, delays.

Cost/efficiency metric focus: The drive to reduce internal manufacturing and transfer risk upstream may reflect a financially-driven management mindset (outsourcing to reduce capital cost) rather than builder/innovation-centric.

Cultural/operational misalignment: Employees and insiders reported that quality was sidelined in favor of schedule/throughput.

MBA fingerprints all over it

“Manage what you can measure”: Manufacturing cost reduction, outsourcing metrics, schedule targets were emphasised; less visible were human/quality/craft/learning metrics.

One-size tool/structure: A standard outsourcing model applied in a context (high-technology aerospace) where emergent adaptation and oversight were critical.

Short-term/financial prioritization: The urgency to deliver and cut cost may have trumped investment in deeper quality and supply‐chain robustness.

The takeaway?

In highly complex, innovation-intensive fields, management strategies that focus too narrowly on cost and metrics (outsourcing, schedule) can produce systemic risk. Robustness and human/quality factors must be embedded from the start.

Enron

How management dropped the ball

Financialization and short‐term metrics: Enron’s compensation and culture were extremely focused on meeting earnings targets, boosting stock price, special‐purpose structures and mark-to-market accounting to show growth.

Ethics/complexity neglect: The financial engineering, opaque off-balance-sheet entities and risk‐taking were demonstrations of tool/metric-based management divorced from long-term value and stakeholder risk.

Oversight failure / culture of measurement: The emphasis on financial benchmarks, growth illusions and internal incentives created a culture where the outcome of the tool-driven management was disastrous.

MBA fingerprints all over it

The obsession with shareholder value and financial metrics at the expense of broader accountability.

A reliance on management models (mark-to-market, SPVs, bonuses) that look good on paper but neglected substance, ethics and stakeholders.

The takeaway?

When management treats metrics and financial engineering as ends in themselves rather than means to serve business purpose, the result can be catastrophic. The “toolbox” enables but does not guarantee good outcomes.

"Hopefully today will serve as a massive wake-up call for AWS"

I wouldn't hold your breath. There will be incident reviews, meetings, assessments, analysis etc. but basically boil down to what can we do to stop this from happening again without actually spending any more money. So no, not hiring fresh talent or retaining that talent already in play, no to radical overhaul of process and knowledge. No to remediation of known issues if it involves expenditure. Instead it will be do more with less. Beat the employees harder, enforce more and more diligence and output from less and less people for the same or less money. Spin it like mad with catchy titles like knowledge sharing, centers of excellence, efficiency improvement initiatives, agile resilience, and continuous operational excellence.

There’ll be shiny PowerPoint decks about empowering ownership and shifting left, while the remaining engineers are shifting caffeine straight into their bloodstream at 3 a.m.

Next quarter, they’ll unveil a bold new policy called Focus Fridays which will be promptly filled with mandatory incident retrospectives. Someone will suggest replacing ancient tooling, only to be told, “We’ll revisit that next fiscal year,” which is code for never.

Then come the internal awards: “Unsung Hero of the Outage” goes to the one poor sod who rebooted the wrong thing but accidentally fixed it.

HR will roll out a “Resilience Recognition” badge on the intranet. This will be marketed with great fanfare and excitement, showcasing how the company truly values it's employees and recognized their contribution because badges are cheap. Leadership will congratulate themselves for “learning from adversity,” and by the time the next blackout happens, they’ll have a snazzy new dashboard to watch it fail in real time along side their investment portfolio dashboard that takes up a greater fraction of their attention.

But don’t worry!!!! There’ll be a T-shirt. “I survived the 2025 AWS outage.” Comes in gray. Just like morale. If it wasn't for the negative impacts on the employees and customers the word Schadenfreude would be very applicable.

And it's a sad indictment on current management practices and in particular the MBA brigade* that this is all by design, acceptable losses on the alter of profit, albeit short-term profit. Efficiency theatre as far as the eye can see.

*Yes, the same people who think Jack Welch was a misunderstood visionary rather than the spiritual father of mass layoffs, short-termism, and shareholder-value human sacrifices. The kind who see burnout as a KPI and chaos as a “scaling opportunity.”

Next they’ll launch a “Transformation Task Force” whose primary transformation will be renaming the same broken process from post-mortem to value realization review. A new acronym, a new logo, and boom, problem solved at a low low cost, honest, the consultants said so. Until the next outage, at which point someone will quote Sun Tzu in Slack.

Larry Ellison's latest craze: Vectorizing all the customers

"The first thing that Oracle did [was take] private data and [make] it accessible to AI models. We took all of our customer data and we vectorized it." Hmmmmmmmm if they took their own data about their customers and waved some dead chickens at it to invoke the appropriate daemons fine. Stupid expensive way to achieve the outcome, but fine.

But if they used customers own data, without consent, permission or notice then that's an intriguing admission from Larry. Which opens up to possibilities, buried in the Oracle T&Cs is permission to do just that, probably phrased as training, teaching or product improvement language and we pinky promise not to share your data outside of Oracle. Or Oracle is taking a Meta approach to data slurping and saying eff it, we own the world, what are you going to do about it, never look in the rearview mirror, all your data are belong to us.

I'm betting it's door number 2 Monty

Cisco: Most companies don't know what they're doing with AI

leaving aside the obvious sales ploy to encourage buying Cisco (or any hardware) AI enabled new stuff I thought the five tips slide was useful, albeit nothing new or original in the last 3 decades or more.

My comments in italic....

Five tips from the Pacesetters

Plan and act with clarity

Pacesetters have an AI strategy. And they act on it. Clear priorities mean less time stuck in pilots and more progress in real-world use cases.​

Ignore the AI component, companies with a clear strategy have clearer proprieties and have more progress.

.

Invest in infrastructure early

Instead of waiting for bottlenecks, they build capacity for scale from the start. That preparation means AI becomes an enabler, not a strain.

Any techie worth their salt builds infrastructure and capabilities not only for today but also for tomorrow wherever possible

.

Treat data as a discipline, not a hurdle

Their data is clean, centralized, and ready to integrate—so AI doesn’t get tripped up by silos or patchwork fixes.​

Takeout the AI bit and yeah very true "Treat data as a discipline, not a hurdle with clean data, centralized, and ready to integrate—so whatever you are building doesn’t get tripped up by silos or patchwork fixes.

.

Lead transformation, not just technology

Change management is built in, which brings people with them. Pacesetters recognize full support turns ambition into action and value.

Yep Change Management is a key component of any transformation or change :)

.

Balance innovation with guardrails

Pacesetters embrace agents, growth, and new use cases—but with governance, security, and monitoring in place. That balance is what keeps value scalable and responsible for them.

Basic 101 practice AI or no AI.

Personally before heading down the AI path at any great length I would be ensuring the following was also in play, but I guess it doesn't sell the hype.

1. Governance & Oversight

- Strategic Alignment

- Accountability Framework

- Policy Integration

- Third-Party Oversight

- Ethical Guardrails

2. Security & Resilience

- Access & Identity Controls

- Model / API Security

- Information Classification

- Cloud / Infrastructure Assurance

- Audit & Monitoring

3. Privacy & Personal Information

- Data Minimization

- Legal Basis & Consent

- Cross-Border Controls

- Privacy Impact Assessment (PIA)

- Retention & Deletion

4. Operational Readiness

- Employee Training

- Incident Response

- Performance Validation

- Change Control

5. Continuous Oversight & Reporting

- Governance Reporting

- Independent Assurance

- Sunset / Exit Planning

Space Shuttle war of words takes off as senator blasts 'woke Smithsonian'

Canada is nuts.

They’ve got everything, massive natural resources, smart people, and the biggest market in the world right next door. And somehow, they’re still managing to look at the UK and say, “Watch this.”

This isn’t about one party. I don’t buy the “new government” line when it’s the same crew from the old one with new titles. Truth is, none of the parties have done much to be proud of. Between provinces fighting Ottawa, pointless trade barriers, and layers of red tape, it’s like being dealt a winning hand and insisting on shooting yourself in the foot twice, just to be sure.

And then there’s the federal transfer circus. Some provinces keep paying the bills while others keep cashing the cheques. Quebec, for example, keeps electricity prices artificially low and leaves a lot of its natural resources in the ground, choices that conveniently make it look “poorer” on paper. The equalization formula only measures what provinces actually collect, not what they could earn, so Quebec’s self-imposed restraint means more transfer money. Call it clever policy or creative accounting, but it’s hard to talk about fairness when the system rewards doing less.

In Western Canada, people are conservative in the real sense. Not the online kind shouting at clouds, not the loudmouths with flags in their truck beds. Sure, those exist but most folks just roll their eyes and get back to work. These are hard-working, decent people. They run small businesses, farms, and trades. They show up, do the job, pay their taxes, and don’t expect anyone to hand them a thing. They’re not obsessed with what others do, not trying to fight culture wars, not looking for attention. They believe in earning your way, minding your own business, and being left alone to live your life.

But they’ve had enough of being talked down to by Ottawa. Every election brings the same promises, the same photo ops, and the same disregard once the votes are counted. And when your Prime Minister rolls by and literally gives you the finger from his train? Yeah, they remember that. Every single one of them even the kids.

https://www.cbc.ca/news/canada/manitoba/justin-trudeau-winnipeg-town-hall-1.3952618

("Why did your dad* give everyone in Western Canada the middle finger?" one student asked the prime minister.) *Pierre Trudeau

https://nationalpost.com/news/politics/a-b-c-museum-says-its-preserved-the-railcar-from-which-pierre-trudeau-gave-the-finger-to-protesters

And to be frank, branding anyone who disagrees with you as MAGA, woke, or whatever label is fashionable this week just dismisses the majority of people on the “other” side as cranks, without ever asking what actually bothers them. The gap between views isn’t a void; it’s an overlap. Probably 80 or 90 percent of it. Focusing on that common ground would be a far better use of everyone’s time.

I'm struggling with he expression "the antique spacecraft". I skived off school to watch the first launch of what was a stunning piece of cutting edge technology. Feel old now :(

BOFH: Recover a database from five years ago? It's as easy as flicking a switch

Ahhh, young Padawan… you do realize, do you not, that there are backups of ‘what was,’ and then there must also be backups of what should have been? Of course you do. It’s rather like that quaint little conundrum the ancients called Schrödinger’s Cat—a creature simultaneously alive, dead, and audited.

For in the minds of the ever-vigilant bean-counters, there exists ‘what was,’ which—naturally—bears no resemblance whatsoever to the tedious reality of ‘what is.’ Restore a backup, and poof! You collapse the waveform into the comforting illusion of ‘what was,’ faithfully frozen in time like a fossilized spreadsheet.

But ah, if one were to… modify… those backups, then balance in the Force might yet be restored. ‘What was’ in the minds of accountants would at last align with ‘what is’ in the backups—and lo, harmony would spread across the ledger, the fiscal midichlorians would sing, and bliss eternal would descend upon Finance.

Former UK prime minister Sunak becomes human Clippy for Microsoft, Anthropic

I think some research on how wealthy people leverage their own "charities" to reduce tax is needed.

https://fightforright.in/how-rich-people-use-charitable-trusts-to-reduce-taxes

Not to mention the "perks" they receive when their "charity" makes a donation to the right recipients.

In the US it is even more naked.

https://inequality.org/article/true-cost-of-billionaire-philanthropy/

Re: Why?

Because once you reach a certain level of wealth it becomes a game/competition and people like that measure their own worth by the numbers. Enough is never enough. And to be fair, I am not sure if I would say no to a token position on an organization paying me a couple of million a year either regardless of my own personal financial circumstances.

But the fact that senior politicians, pretty much anywhere in the western world, can enter the system at X wealth and exit at 10x to 100x wealth should tell us what is really going on. If you are working as a politician for lets say 100k a year and 5 years later you exit political life a millionaire and further get lucrative gigs on the usual suspects boards etc. Doesn't that tell you all you need to know?

SonicWall breach hits every cloud backup customer after 5% claim goes up in smoke

Just to be the pedant in the room, encrypting your backups will stop or slow reading the data but will not stop your encrypted backups being encrypted for ransomware. Thats not what happened here but just in case someone assumes encryption is the magic charm that fixes all evils.

End PSA

The price is wrong! California goes Bob Barker on algorithmic price rigging

Re: all of your data and all of their data

I agree, if I am selling something on Marketplace and I scour the ads to see what other sellers are selling the same thing for and set my price accordingly, that's not price rigging, it is market research. But if I create a tool that allows every other seller of that thing to communicate or indicate a preferred price to sell at and as a result we all set a price as a result then we are price rigging. It can be quite subtle.

If there’s a meeting of the minds to influence prices together, that's price fixing.

Apart from non-public data, which I agree is grounds for breach, I think the other point is any shared collaborative tool that recommends prices also falls foul of the law. Independent decision-making has not occurred.

Zero-day lets nation-state spies cross-examine elite US law firm Williams & Connolly

As a government cyber security person once told me "You will never know the Americans have been in your systems, you'll only know if the Russians have been, if they want you to know and you'll probably spot the Chinese but they are learning at a rapid rate and will soon surpass the Russians" that was 3 years ago.

There is a big difference between securing your IT systems against nation states and securing them against, for want of a better expression, private enterprise hacking. Realistically securing your systems against NSA intrusion is not financially viable. And that's before we even consider that your SaaS etc. are housed in readily available infrastructure in the US or touchable by the US. But that's like worrying about spy satellites taking pictures of you. A bit pointless. But so many of these "breaches" are not relying on nation state backing, basic cyber security hygiene is not observed. Your typical data breach relies on piss poor controls, so once a zero day is found and they penetrate your hard crunchy outer layer, they have relatively easy access internally, can move laterally, escalate permissions, all the good stuff we have seen for years. But with no internal controls, tripwires or monitoring, once in, they run rampant.

I have said it before and I will say it again, proper punitive punishments for Execs and boards would concentrate minds wonderfully. There are some well established frameworks out there to ensure that good hygiene is practiced. If every time a breach occurred, a forensic review by an independent 3rd party took place and if the company was found lacking, fines and prison time were on offering, you would see a remarkable reduction in these types of events.

We have the concept of fiduciary duty, a similar cyber security duty would be useful if properly implemented and with teeth.

Deloitte refunds Aussie gov after AI fabrications slip into $440K welfare report

Re: So….

And 9/10 the report is commissioned in order for those with a lack of critical thinking to say look, we brought in a 3rd party and they recommended x, y and z. Useful in two scenarios, 1) you want extra oomph to get your idea across the line, most organizations suffer from the "no prophet is accepted in his own hometown" syndrome, 2) if or when it goes tits up, you can point at the "expert" 3rd party. Always a useful defense when the board or other reviewing party looks at what happened and says WTF?

London cops unplug iPhone crime ring said to nick 40% of city's mobiles

I'd love to know where they were selling the phones for those prices, I have a drawer full of phones I can happily supply. Pre-wiped ready to go.

Starlink is burning up one or two satellites a day in Earth’s atmosphere

This article needs a rewrite. Not that the facts are in and of themselves incorrect but as you say it jumps all over the place and amalgamates LEO orbits with non LEO orbits, offers no perspective of how the deorbit of 5 satellites a day compares to the natural meteor burnup per day. Is a Kessler tipping point event a huge problem for LEO orbits? Due to atmospheric drag, although conceivably random collisions could eject materials higher up into a longer decaying orbit, current deorbit rules for non functioning satellites is now set at 5 years from the old 25 year rule. (FCC 22-74, IB Docket No. 18-313) Not saying it's a non-issue, just steps have been and are being taken to reduce the likelihood.

A similar google on the makeup of meteors offers the following in order of abundance

Chondrites containing Olivine Pyroxene ( Magnesium Iron silicates) and Iron-Nickel (86% of all meteorites)

Achondrites containing igneous type rock (8%)

Iron mainly Iron Nickel (5%)

Stony Iron 1/2 silicates 1/2 iron nickel (1%)

So a lot of material is burning up daily but it has done so for millions of years and at greater rates than today. But equally not a lot of "exotic metals" which I presume to mean Gold, Titanium, Tungsten, etc but most of the Iridium we have on the planet which is very rare comes from meteors. (ignoring whatever is in the core of the planet during planet formation)

Reading the NASA article the 10% of aerosol particles included copper, aluminium, silver, iron, lead, magnesium, titanium, beryllium, chromium, nickel, zinc and lithium as well as niobium and hafnium which seems to have the NASA article writer puzzled but the referenced article https://www.pnas.org/doi/10.1073/pnas.2313374120 clearly states "Niobium and hafnium are markers for the reentry of some rocket nozzles"

An interesting topic and certainly worth keeping an eye on. I'd also be curious how impactful this is versus one volcanic eruption?

OpenNvidia could become the AI generation's WinTel

Re: Niches

I think you are right, AGI will be some leap in thought or approach, not iterative improvements of LLMs. And for all that the AI hype sucks, there are some use cases where advanced pattern matching is producing useful results. Not necessarily innovative but in materials science, pharma and drugs it is capable of scouring far more information and permutations than a team of humans, highlighting the more likely success paths and giving research teams leads to new and exciting discoveries.

That type of scenario is, IMHO, ideal usage of LLMs. Curated data fed to a pattern matcher, flags interesting patterns or areas worth exploring, human subject matter experts review the findings, whittle them down to the most likely results and then test the results. It works as a more general expert system in that scenario, and the lack of polluted data sources keeps halucination to a minimum and the results are verified by humans for more exploration.

I think of this as like the autopilot in a plane, sure it could be left to fly the plane on its own but you don't. The pilots use the autopilot to reduce the drudgery and monotonous operations and are the ultimate decision maker if things get out of control. Well, unless you are Boeing and override the pilot inputs because your algorithm knows best and then don't tell the pilots this is a possibility, see MCAS :(

Re: Dot com bust

Some thoughts on the dot.com bubble comparisons

First mover advantage, Nvidia/OpenAI remind me of Netscape, Yahoo, Cisco and Nortel, only one of those 4 is really left standing

Money, dot.com was at a time of booming economy, lots of easy money at low rates, AI is at a time of less easy money (still relatively easy) but higher interest rates. Also dot.com companies had very little revenue, Nvidia posted 26Bn in Q2

Beneficiaries, dot.com was thousands of startups splitting the money, AI is maybe 10 and they hold the majority of the patents too.

Infrastructure, in dot.com bandwidth and servers were toy like in comparison to today, AI is more about power issues.

Assuming that AI pops it's bubble like dot.com we still got lots of fiber and datacenters that went on to service the post dot.com internet. If AI goes the same way there is going to be a lot of compute available at relatively cheap cost.

And post dot.com left companies like Amazon and Google standing tall, they at the time weren't exciting unlike pets.com and Webvan. Remember those? Post AI, identifying the likely winners today, might need a crystal ball, could be an interesting exercise. I'd be looking at utility companies, Synopsys and Cadence FWIW

To digital natives, Microsoft's IT stack makes Google's look like a model of sanity

Re: Blast from the past

And business embedding everything into Domino. I made some serious coin untangling that particular rats nest of kludges over the years. Domino really was the original "to a man with a hammer everything looks like a nail" syndrome.

Re: His mistake is in thinking it was designed...

Don’t forget that Microsoft Small Business Server was in the mix too, and many saw SharePoint as taking the same ethos and scaling it up for the enterprise. The challenge has always been that SharePoint is a bit of a question mark. What exactly is it supposed to be? It’s not a great document management system, it’s not a great file share… in fact, I could list a thousand things it isn’t. But you get the picture.

From day one, you had SharePoint Team Services for individuals and small teams, and SharePoint Portal Server for enterprises. The latter was pitched as a central portal for storing files, managing versions, and indexing content across file shares and Exchange public folders.

If I had to summarize what SharePoint was meant to be at the beginning, it would be document management:

Central repository: A place to store and organize documents with versioning, check-in/check-out, and metadata, so files weren’t just random Outlook attachments.

Intranet/Portal creation: A “single front door” to corporate information.

Search across systems: Indexing capabilities to crawl file shares and Exchange public folders.

Basic collaboration: Team sites for calendars, task lists, announcements, and links.

In short, SharePoint was originally meant to solve this problem:

“We have thousands of files scattered across network drives and email. Nobody knows what’s current, where to find it, or who owns it. Let’s create a central, web-based system where people can store, search, and collaborate on documents.”

But like most things it took the original complexity it was trying to solve internalized it and then layered it's own complexity on top. This was against a backdrop of iterations of outlook, exchange, every flavour of groups you can think of, tied into teams etc to give you what you have today. It wasn't designed on a clean sheet of paper for a greenfield site, and it shows,

Intern had no idea what not to do, so nearly mangled a mainframe

Ah yes the good old tighten it to just a fraction before it starts to feel like letting go. With enough practice you develop a clairvoyant like sense for they yield strength of a particular nut/bolt combination. You may complain that it is overtightened and under stress will fail, and you may well be right, but to be honest sometimes tightening to a torque and then add 90 degrees makes me feel a lot worse about likely failure.

UK and US security agencies order urgent fixes as Cisco firewall bugs exploited in wild

My quote was directly from the article. (2nd paragraph, 1st line)

Are you saying that Cisco had previously issued patches to address the exploit patched on Thursday evening? Or are you saying this was known since way earlier without a patch until Thursday?

"The networking giant has also admitted that it knew these flaws were being exploited as far back as May, when government incident responders called it in to help investigate intrusions on ASA 5500-X firewalls. Attackers were already dropping implants, running commands, and siphoning data"

Which government out of curiosity, the article doesn't make that clear. The cynic in me says whichever government it was, wanted to use it themselves for while before raising any flags.

Regardless, 4 months between spotting an exploit actively in use in the wild and patching is damming for Cisco. Given their gear is used by a lot of agencies, you would hope a swift foot up the rear-end is winging it's way to Cisco.

But if you don't want to use Cisco, who else is there? Palo Alta? Fortinet? Juniper?

Personally in order of speed for patching I'd rank them in order of Juniper, Palo Alta, Fortinet then Cisco, but that just my own personal back of an envelope opinion.

Many employees are using AI to create 'workslop,' Stanford study says

Re: If you have a shitty manager

Upvote for the on point diatribe. :)

Hardware inspector fired for spotting an error he wasn't trained to find

Having a bit of a flashback to my first "proper" job building prototype PTH style PCBs. I could solder very well, could read and understand resistor codes, knew which end of a chip had pin one etc. etc. What I failed to realize on day one was that, although at casual glance tantalum capacitors look like ceramic capacitors, they are polarized and yours truly got approximately 50% of them placed and soldered backwards. That was a lot of magic smoke on first power up.

Re: Maybe they wanted someone incompetent to "inspect"?

Way back in time when Dunlop had a plant in Cork, Ireland rumour was the Friday batch of tires would be overcooked before a long weekend to ensure everyone got some nice long weekend overtime.

The sweetest slice of Pi: Raspberry Pi 500+ sports mechanical keys, 16GB, and built-in SSD

Re: a better keyboard...

True but my God it was so much better than the ZX81 keyboard. Mind you I was still hunting and pecking at that time so the Spectrum keyboard wasn't the bottleneck in my productivity :)

Workers: Yes, RTO makes sense. No, we’re not going to do it

Re: "unbossing"

Yep "technical career advancement track" was one I saw and not 40 years ago either. You have a technically competent person, willing, intelligent and useful. But they either don't want to manage people or you wouldn't want them managing people. So you advance them up the technical track, put them in virtual teams with a decent BA and PM interfacing them to the general populous and good things happen. Occasionally you get the whiny "I manage x people and he/she manages none, how come they are on the same pay scale?" to which I would reply "Because they earn that pay by being dammed good at their job, are a lot harder to replace than you due to specific skills versus general"

At the risk of pigeon holing people generally these are 25+ year veterans, more often than not male, zero people skills (by choice), and take their role and tasks very seriously. Usually a person you have on speed dial for SHTF problems. That same person who is in day to day interactions a pain in the back side to deal with will, after doing a fair bit of I Told You So, move heaven and earth to fix, build or remediate whatever SHTF doodoo you are knee deep in. That's why you keep them fat and happy with career (salary actually, typically they don't care one iota about titles) advancement.

AI coding hype overblown, Bain shrugs

Re: No training sounds about right

The MBAs spout "synergies" and "efficiency" and "cross functionality" and promise huge RoI, increasing profit margins and whatever else will sell. And when, not if, it all comes crashing down, it isn't because the idea was bad, oh no, you didn't implement it properly. Or maybe you didn't do enough SWOT analysis :)

Or maybe, just maybe, it was a dumb idea to start with, but hey the consultancies got their invoice paid, tag your organizations logo onto their PowerPoint of other companies they have "helped" and move onto the next sucker. And the powers that be can point to "expert" opinion to show they weren't at fault which I suspect is the primary reason consultancies get hired in the first place.

Still it does reinforce the notion that in a gold rush the people selling shovels, supplies and prostitutes make the most money.

Stop runaway AI before it's too late, experts beg the UN

Re: This is like when they all got in a flap about blockchain

Depends on the 3d printing you are referring to. I have a Bambu Labs 3d printer that has been fantastic for printing anything from a replacement collar for a 1970's Dodge fuel sender to brackets for solar lights. Once ABS and PETG became easily usable the use cases expanded dramatically. If you mean printing houses, buildings etc. yeah they look horrendous. But I started life in engineering and learned to draw using pen, paper, set squares and a T-Square so jumping to parametric CAD was surprisingly easy. If you can't create your own STLs, relying on Thingverse et al would be a depressing experience.

Metal 3d printing is surprisingly good although I can't afford anything that would be under any pressure or load but a colleague drafted a complete oil to water heat exchanger for a BMW bike, had some Chinese company like PCBWay make it in stainless and it's still going 2 years later although to be fair the amount of hand polishing needed would have put me off. If Rolls Royce can use 3d printed parts in turbofans it's probably a cost versus practicality issue.

Blockchain? Yeah that really was a solution looking for a problem. I still remember a CEO asking me "Are we using Blockchain? Everyone else is using it."

Fundamentally this boils down to do you think AI, LLM, algorithms etc will get to a point where they are capable, whether designed to or not, of making decisions and do things that are bad for humanity. Arguably they already have reached that point and did so a long time ago. "the computer says NO!" meme was a thing long before memes.

I would split this into a few scenarios

Scenario 1 automated systems given access to systems and things to make decisions or take actions that are undesirable. Long standing problem, with long standing guardrails in place to prevent bad things from happening since about the mid 70's. We didn't let computers launch missiles for a good reason, not that it would deliberately, with malice and forethought launch, but rather through bad programming or edge case make a bad decision that would be catastrophic. But now we have the tech bro culture impersonating every evil villain in books and movies doing what they think is best for them, their company, humanity, take your pick, and pressing on regardless. That's a regulation issue and any politician who can't see this coming over the horizon is an idiot, which probably means we are f**ked.

Scenario 2, sure LLMs etc are dumb* stochastic algorithms that understand nothing in the true sense of understanding, but get better and better at mimicking human intelligence until to an outside observer they are indistinguishable from a human, albeit an low IQ human. Think of a parrot but with access to everything ever written and a better programmatic ordering of responses. We are progressing down this path at a rapid rate. Lots of issues to deal with there and doom scroll on Facebook for 30 minutes and tell me which is worse, collective social media or ChatGPT? Expect the money to follow the most lucrative way to leverage these for maximum profit with scant regard for anyone's quality of life.

Scenario 3, the scary one, while scenario 1 and 2 continue to evolve (or devolve depending on your point of view) presume we make another leap of logic and we invent an approach that results in intelligence indistinguishable in every way from true intelligence. And it evolves at a pace and scale that leaves us in the dust and we are effectively dealing with a God. Or a singularity event if you want to use the buzzword. That needs controls and protection. Is it likely? Don't know. But given the rapid scale of advancement we have seen in the last few years if, and I admit it is a huge if, if it happens, we won't know until it has already happened and by then it is too late. So yeah let's game that out, put controls in place and take some steps, just in case.

I would also argue we already deal with large AI's and have done so for 100 of years, they are called corporations. Filled with hundreds or even thousands of minds, all working towards a goal. Every corporation has a collective intelligence far superior to any one human. And we see how corporations behave. Without controls and regulations they consume resources, cannibalize other similar creatures, operate with limited regard for any human and left unchecked would destroy worlds. And you can go back a few hundred years and look at the East India Tea company, Hudson Bay, robber barons for good examples of unchecked avarice to see how that works out. And that's an intelligence created from lots of human intelligences. Imagine similar but now with no internal human intelligence, operating at several orders of magnitude faster and more ruthlessly.

So do I think some thought should be put into this, definitely yes. It's a risk vs probability matrix, if the risk is low but the impact is terrifying then you shouldn't ignore it.

All of this to say, while I don't disagree with those saying it's all hype, buzzwords and in reality LLM's are nowhere close to intelligence, discounting the potential for a quantum leap, be it evolutionary or someone makes a startling discovery, is shortsighted and having open discussions about the what if's isn't a bad thing.

*dumb meaning lacking understanding, recognition of actions versus consequences or any semblance of empathy or morals.