Posts by Pete 2 3751 publicly visible posts • joined 10 Jun 2009
The problem is that satellite orbits are determined on a "first come, first served" basis. There is no method of rationing orbits by country or any other factor. If you can loft 10,000 satellites before the other guy, it is their responsibility to plan around what is already up there, not yours to stay within a specified number or location.
Those outfits that get government contracts are experts at pushing up costs. Primarily because government departments have no incentive to keep them low - it's not as if it's their money!
Ultimately all government initiatives cost as much as the Treasury is willing to pay. This number bears no connection to original estimates. I have a sneaking suspicion that if the suppliers were developing this solution for their own benefit, it would come in on time and under budget.
> rejected the £1.8 billion cost forecast
No matter what number they choose / guess, we all know it will exceed that by an unimaginably high factor.
Then after overspending by an eye-watering amount, and taking considerably longer than anyone could rationally imagine, it will be cancelled when 90% complete.
> booking double time for half days on Saturday's and Sunday's while doing no work as they'd done it during the week.
I saw the opposite while employed¹ by an ex-nationalised industry. Some staff members, mostly those without partners or family (or maybe ones they wished to avoid) spent their Mon-Fri time generally just chatting, reading the paper, drinking coffee and taking long lunches. Then they would come in at weekends on time and a half to do what should have been done during the week.
[1] as a contractor
> Sounds like a lot of work to figure out what's true and what's not,
In general, the best way to get less biased "news" about a country's politics (though why you should want to is also highly dubious). is to see what sources outside that country are saying
Many times, things that are splashed across "news" sites within a country are not mentioned at all, outside of it.
> All they do is predict tokens from training data
Which is essentially all that humans do, most of the time. Except we call it "schooling" and "knowledge".
There is precious little that humans ever invent. Which is hardly surprising as society demands behaviours and actions (and possibly thought) be constrained by social norms, laws, rules and peer pressure.
And as the title reminds us, often when individuals encounter something that Isn't quite right, we auto-correct it to fit what prior experience from our "training data" predicts.
> a sharp increase in workforce overcapacity by 2028 as productivity gains accelerate,
But that doesn't lead to redundancies. If productivity keeps pace there is no reason to cut jobs. It only needs new roles to appear. Mostly ones that contain the letters "AI" in the job description.
> Sigma2 aims to reuse waste heat generated by the liquid cooled Cray EX4000 compute blades to warm water for local salmon farms
#include <spawn.h>
int posix_spawn(pid_t *restrict pid, const char *restrict path,
const posix_spawn_file_actions_t *restrict file_actions,
const posix_spawnattr_t *restrict attrp,
char *const argv[restrict],
char *const envp[restrict]);
> developers are the "ideal victims" because their machines "contain the keys to the kingdom: production credentials, crypto wallets, client data."
So the idea is that a potential employer sends a package of code to a potential employee and expects them to run it on their current employers production machines? Or at the very least, production-adjacent machines.
One might question the security arrangements that permit unaccredited software to be downloaded onto such systems in the first place. Maybe what needs to happen is that all such incoming communications get "AI"d to flag these goings on.
Sure, some might raise issues of privacy. But nobody working on production systems would realistically expect those machines to have the same status as personal kit used for private (i.e. non-company) activities?
> provide "air-gapped" sovereign cloud services and AI in "completely disconnected, highly secure environments."
I have one of those on my desk. It is completely disconnected and highly secure. Nothing goes in and nothing comes out. Given that it is (or claims to be) a "sovereign"AI cloud, it is surprisingly small and lightweight. I did wonder why it was cheap, though.
> "It was great knowing you Arduino, RIP. Hello, RP2040 and ESP32
Twenty years ago when Arduinos arrived, they were revolutionary. These days the hardware is increasingly irrelevant. The legacy that Arduino has is the huge amount of user-contributed code and it seems to me that is the only value associated with the boards now.
> We will work with NHS England colleagues to co-develop a roadmap that establishes the criteria for value-based adoption and identifies the point at which the FDP adoption is in the best interests of the GM population."
.
You do have to wonder why NHS England didn't simply take the Greater Manchester system and adopt it nationwide?
Maybe it was too "northern" for a London based decision making body?
One place I worked, my job description could be called "anti-sales".
When new equipment was needed, and with an IT budget of over £100 million a year, that was all the time, I was tasked with working out what specifications were actually required. That was to counter the suppliers sales teams who considered the company an easy target, since the people signing off on new kit knew didily-squat about MIPs, gigabytes or bandwidth. The closest they got to technical knowledge was "oooh, shiny".
After one carefully prepared, numerically sound, counter-proposal the question of software licensing came up. I told the PHB how many were needed now and what the future would require. However, the wily salesperson had slipped in a bunch of extra licenses. Much like a dishonest waiter will "accidentally" add another round of drinks to the bill.
It turned out that bossy-kins had assumed both numbers were the same and put the order in, with the inflated license numbers. When someone questioned the extra cost, manager-person replied "oh well, it's only £60k" and the audit department was satisfied with that.
> Make sense of Microsoft licensing
You will give us money as and when required
The amount will be notified as we see fit
You will only run applications we like
They aren't bugs, they are features
None of our products come with any guarantees that they will do what you think they should
We reserve the right to change any API, at any time for any reason
Any key press or mouse movement may be accompanied by a popup advertisement
Your data is now our data
Most companies are willing to pay for software support and licences. Many will avoid software packages that do not have a sound commercial basis. The feeling being that provides them with a degree of security and indemnity. Security that a single vital FOSS author won't just flounce off, discover sex, die or decide they dislike certain companies and single-endedly change the terms they distribute their contributions under.
Plus, a support contract will oblige the supporter to provide fixes within a given time. Something that risk-averse C-levels value highly. Especially when they have cut their IT budget to the bone and don't have any competent programmers to fix things, themselves.
> guaranteeing at least 15 years of software updates.
But this does not stipulate they must be free updates. So all that vendors have to do is ramp up the cost of updates over time. Until the cost of updating exceeds the price of a new box.
The software vendors can then say "It's not us, guv. Look! We're doing what the law requires. They are throwing away working kit"
What is worse is that once the model makes paid-for upgrades and fixes more common,then it soon becomes the norm. For everyone.
> AI-assisted developers produced three to four times more code than their unassisted peers, but also generated ten times more security issues.
Possibly forty years ago, I came across a cartoon in one of the IT industry monthlies. It showed two techies in a machine room surrounded by huge cabinets of computing equipment (stuff that would now all fit in a matchbox).
One was saying to the other "with this new computer the boss is able to mess up a project in half his usual time". It seems that little has changed.
> consumed roughly 45 milliliters of water
Was that water really consumed, i.e. destroyed? Or was it simply changed from one state (liquid) to another (vapour)?
If that water actually did cease to exist, either as molecules or their component atoms, I can see that would be quite a problem. However, if the water just went to join it's own "cloud" then all that has been consumed is the energy needed to transport it's replacement. While that is significant when scaled up to gigawatts of cooling in a data centre, the actual problem is different and has different solutions.
> 45 percent of IT workloads are still operating on-premises in corporate datacenters, with another 16 percent in colocation. Only 11 percent are in public cloud infrastructure, although another 10 percent are listed as hosted private cloud, and six percent is made up of software-as-a-service.
45+16+11+10+6=88.
If I was doing capacity planning I would be rather concerned about the quality of the data I was given.
One office I worked in sported the largest conference room among all the various locations the company had. As such it was often used by the higher-ups for their strategic meetings. Even though there were only eight of them (in a room that could hold 40) I suppose there was prestige to sitting around a big table.
What they were unaware of is that the inside wall of the conference room was not soundly (I use the word advisedly) attached to the rest of the building and the occupant of the desk where that plasterboard wall met the exterior wall would often find themselves in a position to hear what was being said inside, in assumed confidence.
There are thought to be some billions of different microbe species. Of which a few thousand have been discovered, classified and named.
Rather than evolving in space, the laws of probability say its millions of times more likely to have just hitched a ride. I fully expect that if the chinese look hard enough, they find find many more. As would anyone who cares to inspect the ISS.
> The ML admin articulates, designs, plans, executes, and monitors the emergent large language model lifecycle
Any more buzzwords we could stuff into this sentence?
I can see the need for an MLA, just as I understand the need for a DBA, network admin and all the other specialised roles. The question that comes to mind is why this would have to be a person. Isn't the entity best placed to monitor LLMs and all the rest of it just another LLM.
In fact I can foresee an entire shadow organisation appearing within IT companies. One that is made up entirely of AIs, agents, MLAs and a few yet-to-be-created roles in AI management. I just wonder how much of their time will be spent in pointless machine to machine meetings
That should slow 'em down to the point of uselessness.
> So I'm sure forcing people back into the office will kill that attitude
Was happening long before Covid.
When the location of the 2012 Olympics was announced (in 2005) one of my co-irkers essentially stopped doing their paid-for job for several weeks. Instead, while using the company offices as a base - so lax / incompetent was t'management, they spent all their days in property speculation.
It was not because they hated their job - they did so little actual work there was nothing to hate. It was simply because they could.
> an unidentified "data partner" ... infer their place of employment.
So really this is just a technique of espionage. As well as inferring a target's person's place of work, the very same process can (and therefore does) infer many other things, too.
Some of which might even be correct
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
