The Register Home Page

* Posts by matjaggard

405 publicly visible posts • joined 22 Apr 2015

Page:

Home Office cloud services contract may net AWS £1B

matjaggard

Re: Why?

I'd agree the home office would be best hosting some things themselves but clouds are great for expanding quickly, handling spikes and doing experiments. That applies just as much to government as anyone else. A government cloud that can handle much of that across all central and local government departments might be a good idea but they'd only go and outsource it to some awful consultancy.

matjaggard

Value Added Tax

"passes £1 billion once VAT is included"

is it wrong that I find the idea of the home office adding value laughable?

Node.js creator ditches Deno to work on open source Cloudflare Workers

matjaggard

Are *all* the commenters on here retired?

Anthropic decides to support OpenAI's markdown instructions spec

matjaggard

I'm also glad for this fix but you could always just put @AGENTS.md in your CLAUDE.md file and it did the same thing as the symlink (even without extra token usage) plus you can then have some items that are Claude specific in there too.

Casio decides it's about time the simple digital watch got a little smarter

matjaggard

Re: Just the day and date here.

He did say he does have the day of the month, just not which month.

I on the other hand *do* forget the month, and even the year from time to time but I accept I'm exceptional in the worst possible way.

Be a hater all you want, AI's here to stay

matjaggard

Re: Are you sure about that?

An LLM is not a state machine. It doesn't have a finite number of states, it doesn't have a definable state function, it is not deterministic.

matjaggard

Re: 95% of AI isn't financially viable.

There's so much incorrect information in this thread.

For example most of what's labelled as AI in the last couple of years is LLMs so absolutely not "old-fashioned data processing" - that mislabeling was a few years ago.

Also we absolutely do know the costs of running LLMs - big 3 cloud providers post figures, and their pricing is pretty close to the correct amount. Anthropic and OpenAI give massive discounts for per-seat pricing but API prices are about right. Running open weight models on your own GPU comes to a similar cost if it's amortized over 6 or so years and fully used.

AI will absolutely not be fading away, as you'd know if you worked with it.

matjaggard

Re: I just don't want AI shoved down our throats

This isn't quite true, we're being chased for our next AI features all the time, and not just from managers, from other companies and the NHS. End staff might not want it, but they need to tell the procurement teams if that's the case.

matjaggard

I mean you're welcome to your opinion but it's factually inaccurate. Having said that, companies who lay off staff because their jobs are being automated are shortsighted and should focus on the new things we can do - that's where longtime success lies. Keep your employees and do more, not just make a fast buck by laying them off.

matjaggard

That's not really true. You can use AI to make a quick and dirty solution, sure. But you can also use it to make reliable and well designed software and to maintain that software.

For me personally the primary things it's done are make me able it apply my years of experience in more areas - for example I rarely care about if I know a specific language anymore - and to speed up delivery.

Don't get me wrong, it has downsides - whilst I still get the satisfaction of delivering good value for customers, I do miss the act of coding - I also worry about the next senior developers as the article mentioned, but overall it's a net positive to me.

AMD’s Ryzen AI Halo makes local AI look easy, but at $4K, easy doesn't come cheap

matjaggard

Is it worth it?

When discussing the price, why wasn't this compared against cloud offerings of open weights models? In my experience running a model in one of the major or local cloud providers allows you to specify your data stays in your own country for a fraction of the cost, even spread over lots of years - and who knows how long this hardware will last / be useful for.

Feds freaked over Fable 5 after simple 'fix this code' prompt, not jailbreak, says researcher

matjaggard

Re: Stunt

@cyberdemon I was with you until you said that it's useless for defenders. It is absolutely not - it can find an fix almost all well known classes of bugs and that is enough for most of the world. Very few security incidents are caused by something novel. It's also got very good recently at not breaking other code now by using semantic search and abstract source trees.

Brit workers waste nearly six hours a week 'botsitting'

matjaggard

I mean kind of. Training it appears.

A Russian speaker and jailbroken Gemini went on a hacking spree and emptied at least one MAGA victim's crypto wallets

matjaggard

Re: So Now Reg Writers are Quoting "Security Professionals" without Challenging Lies

I totally agree with the "come off your high horse" comment. "Reasoning" is a perfectly valid technical term for when an LLM talks itself through a prompt before responding back to the user. Is this human reasoning or human intelligence - no, and nobody is claiming it is - but re-using human terms for what a machine does is both perfectly reasonable and as old as computers themselves. They can't "calculate" like a human, or "analyse" like a human or "decide" like a human but of course with have decision trees in programs, analysis of data, even in simple if/else code, calculations.

matjaggard

Re: Only problem I see

Maybe he's hoping to be spotted as a good hacker when his conscription papers arrive.

matjaggard

Re: Jailbreaking?

I am absolutely *not* planning on typing `vi GEMINI.markdown` - leave the two character extensions alone.

matjaggard

Re: The campaign targeted the QAnon and MAGA communities

That's not a great reason to avoid crypto. "Real money" doesn't have any basis in real value! Best to avoid it due to the types of people who are pushing it, the corruption, the reliance on ever increasing amounts of compute power (in some cases), the lack of backing and guarantees by your nation state, but not due to the basis which is missing from all currency.

matjaggard

Re: So Now Reg Writers are Quoting "Security Professionals" without Challenging Lies

Well that's just not true, there are loads of profitable uses. Whether profits should be the real aim for a change this big is debatable but whether it's profitable is absolutely not.

NHS England confirms: Palantir staff can access patient data

matjaggard

They're literally a spying company

do we really think NHS England have got even the slightest assurance that this data will remain in the UK?

Cloudflare to fire 1,100 staff whose jobs just aren’t AI enough

matjaggard

Really, there's no need to mention that you're old, you're in the comments section of The Register!

Claude hitches ride on SpaceX's datacenter capacity

matjaggard

Re: Self inflicted wounds

Thanks. I took the opportunity whilst reading those to click on a few adverts for Starlink

Inference is giving AI chip startups a second chance to make their mark

matjaggard

Surely there's a better way?

Surely there's a better way to refer to autumn 2025 than "last northern hemisphere fall"

matjaggard

Re: Big words the author doesn't seem to understand.

Language is allowed to change, be imprecise and have two opposing definitions. Annoying I agree but scientific terms are particularly bad at being used by others and taking on new meanings.

Anthropic admits it dumbed down Claude when trying to make it smarter

matjaggard

Actually I don't think this is about them trying to get money. I mean obviously that will come but this is because they're expanding customers and customer usage faster than they can build compute.

matjaggard

Re: Cross the streams?

That's true but shows no signs currently of happening. LLMs are getting better and more capable of executing tools all the time.

Anthropic squeezes enterprises by ejecting bundled tokens from seat deal

matjaggard

Good luck buying suitable hardware!

Claude is getting worse, according to Claude

matjaggard

Re: Gosh, who knew ?

GitHub. Currently at zero 9s https://mrshu.github.io/github-statuses/

matjaggard

Re: I cannot even begin to count the problems with that approach

You're quite right that this is not a good way of getting quality data - mostly because of the increase in Claude users during that time. Also as time goes on the people using Claude will likely be less technical (at least as far as my company goes anyway).

NHS staff resist using Palantir software

matjaggard

Re: Wouldn't matter who it was

That's a very cynical comment. I mean Palantir are likely the worst company in the entire world for handling British medical data, but you're right its a tough crowd to win over.

Anthropic tweaks timed usage limits to discourage Claude demand during peak hours

matjaggard

Re: Of course they don't tell people how to plan for it

Of course the prices are going to increase, that's a given, but I think you're underestimating the users of Claude. I think the majority will know what a token is and whilst they probably don't, they could estimate how many of them they're using. Anthropic are playing a slightly dangerous game by making their costs so opaque - people are not stupid, they will build things to replace Code, Co-work, etc. if the pricing goes too high.

matjaggard

Re: Words

I can't even tell if you're joking. You know most Claude usage is for work right?!

UK watchdog eyes Meta's smart glasses after workers say they 'see everything'

matjaggard

Re: describing the allegations as "concerning."

Well aren't you all pleasant people!

A lot of people using these are blind - they have small speakers in the arms and so a bind person can suddenly be able to read signs, know what's around them, call a family member if they feel confused or concerned.

But instead of reading a story like this and thinking "these tech companies should stick to the law" you just call for them to be illegal or worse, "euthanise" the users.

You should be ashamed of yourselves.

Am I surprised the tech companies need a wrap on the knuckles? not at all but that doesn't mean we should ban innovation.

Anthropic tries to hide Claude's AI actions. Devs hate it

matjaggard

No, they've been totally clear that it can and will make mistakes.

matjaggard

Re: "The session just burns through my token quota."

That fits with the last change they did too - removing the cost summary when quitting the tool

Java developers want container security, just not the job that comes with it

matjaggard

Reads headline:

*I wonder if BellSoft by any chance sell hardened docker images*

Gosh, they do!

In-house techies fixed faults before outsourced help even noticed they'd happened

matjaggard

Almost believable

I was with you right up the the contractor waiving a fee - took the story from "incredibly likely" to "apocryphal".

Lawyer's 6-year-old son uses AI to build copyright infringement generator

matjaggard

I think you're wrong, I think despite us not paying them like they'd get in the private sector lawmakers are at least 90% *trying* to do the right thing but sadly the odds are stacked against that outcome. They don't have the tools they need to make the right laws so they have to take advice and that advice generally comes from people with enough money to get to lobby them.

The whole funding model, choice of politicians and lobbying rules all need to change to fix this. Thankfully here in the UK we do the bare minimum of limiting political spending but it needs to go so much further.

'Vibe coding' named Word of the Year. Developers everywhere faceplant

matjaggard

Re: Having tried vibe coding

This is not vibe coding. This is an expert using a new tool like Clause or Copilot.

Having a AI write code for someone who fully understands and problem domain but can't code, and is willing to sit and chat to a tool for hours can produce something pretty good. That's Vibe Coding. Outputs can be functional, you can start a business with it before needing to re-architect at some point. The code is actually pretty well laid out and well written if you can ignore the excess of comments and emojis.

Our current issue is how to deal with this in the longer term. Do you have a moment when it's thrown over to proper devs? We've learned from old methods of testing and hours of Jira ticket writing that throwing things over the wall to the next team doesn't work well but how do we handle this if not?

UK government on the lookout for bargain-priced CTO

matjaggard

I'm very tempted to down vote just because I wish this wasn't true.

Sneaky Mermaid attack in Microsoft 365 Copilot steals data

matjaggard

Re: Not a bug, a feature

Either way the message is clear - it's better to exploit or sell the vulnerability than tell Microsoft.

Who gets a Mac at work? Here's how companies decide

matjaggard

This is why my boss chose Macs for my whole department at Barclays - basically IT were less good at locking them down than Windows machines which meant an order of magnitude less time fighting against the Request system to get the software you needed.

RubyGems maintainer quits after Ruby Central takes control of project

matjaggard

Re: While we're here, whatever happened to Ruby on Rails anyway?

My company uses it and we all hate it. It's very fast to get something working and very slow to keep it working because of all the magic it does. Things just magically get changed from camel case to snake case for example, types are impossible to follow and can change as you use them due to the "becomes" keyword, and Robocop is more Nazi than police by default.

Sorry, but DeepSeek didn’t really train its flagship model for $294,000

matjaggard

I think you've all missed the point here. The discussion is about how much compute was used and how much that would cost. The human input is ignored for this discussion and hardware costs aren't massively different between locations because shipping is cheap.

Microsoft open-sources the 6502 BASIC coded by Bill Gates himself

matjaggard

I found a mistake

I found a mistake on line 119 of the main assembly file.

Older developers are down with the vibe coding vibe

matjaggard

I think you might have a touch of confirmation bias in your reading.

I also dislike black boxes in my code but I don't mind LLMs because I know how they work and from experience where they don't. Probabilistic output is not quite the same as a black box.

I've not found it any harder going back to code I've accepted from an LLM vs code I wrote myself. I just don't accept anything bad from the LLM.

AI doesn't solve all problems, probably not more than 5-10% but that doesn't make it useless. It makes it a tool that you can learn to use and those who do will likely succeed more than those who don't in the long term.

matjaggard

Re: But...isn't it all just bollocks? ... Yes it is !!!!

Your logical process there has a mistake. Someone competent enough to spot a mistake might well want to use AI for a number of reasons. For example LLMs are incredibly good at finding salient points in masses of data - documents, code, etc. Given the correct task it can save huge amounts of time and for many problems, checking it is trivial.

Users of PostgreSQL in the cloud say the uptime just ain't up to it

matjaggard

Re: I still fail to see the point

Well in my experience it's because it just works. My app has uptime of 5 nines and I've never had to do anything much with the DB except pay through the nose.

Post Office and Fujitsu execs 'should have known' Horizon IT system was flawed

matjaggard

Individuals

Why are people focused only on individuals who should be dealt with whilst Fujitsu are STILL winning government contracts??! I'm not saying individuals don't need to face justice (but not violence - come on people) but we can't ignore the continuing decisions to pay more and more to Fujitsu.

Former and current Microsofties react to the latest round of layoffs

matjaggard

Re: nobody asked for

I think the use of the phrase "nobody asked for" in saying what Microsoft are doing is to miss how business works. Tech businesses that *just* do what people ask for will fail because anyone can listen to people and produce something that fits their stated needs. The difficulty is fitting their actual needs or doing something brand new that people haven't thought of.

That's where AI is currently and it will change our entire world - especially for developers who need to understand and eventually embrace it or may as well quit.

UK dumps £2.5 billion into fusion pipe dream that's already cost millions

matjaggard

Re: Rewrite needed

The bigger question which isn't even asked here is who will own this? Will we as a nation own (at least some of) the resulting Intellectual Property and/or facilities and companies that this could create?

Page: