HNHacker News
TopNewBestAskShowJobs

xyzzy123

7,319 karma · joined September 21, 2012

contact zxcdotmx@gmail.com
submissionscomments
xyzzy123··on Dyson settles forced labour suit in landmark UK case
Why not push it all the way to the consumer? Why shouldn't you be liable if you buy a wrench, but actually the worker who made it was mistreated? That would make people think twice before buying products of unknown provenance and supporting slavery.
xyzzy123··on What Claude Code chooses
So I wonder about sharable skills? Like if it's a problem that lots of people have, I find the base model knows about it already.

But how to do things in your environment? The conventions your team follow? Super useful but not very shareable.

Whats left over between those extremes does not seem to be big enough to build an ecosystem around.

Final problem, it seems difficult to monetise what is effectively a repo of llm generated text files.

xyzzy123··on The Slow Death of the Power User
My point was meant to be a bit more subtle; that we all engage with abstractions of reality. Whenever I'm tempted to complain about people not knowing stuff, I pause and think about how little I really know.

In my grandfather's peer group, you were lacking essential knowledge if you couldn't rebuild an engine.

xyzzy123··on The Slow Death of the Power User
Most people, when they turn the tap on, they don't know where the water comes from. Try asking someone "What is the physical principle that makes the water come out of the tap? How do they make it come out?" you might be surprised how many people don't know and most importantly don't care.

The water comes out. The water has always come out, every time, so it's not really a thing worth investigating. Like the sunrise.

In many many domains I am that person.

If a person doesn't know (except in the vaguest terms) where their water comes from, where their poo goes when they flush, where their food comes from (the supermarket!), or the energy that heats their home... what do they really know? Most of us know very little about the concrete networks and systems that keep us alive.

But this is what civilisation is.

xyzzy123··on Aesthetics of single threading
I love to single thread but nobody else seems to. A typical situation would be making a sandwich for one kid while two others are trying to talk to me at the same time, each rising in volume to cut through the noise of the other. Partner explains in roundabout way that something is needed tomorrow. I wonder what to do with that information and wish it had been communicated in fewer words while making the sandwich. Then the phone rings.

A lot of the time work has this character also.

xyzzy123··on CXMT has been offering DDR4 chips at about half the prevailing market rate
Agree, worth analysing what is genuinely commodity.

There are more elements to it though which can be sort of hard to explain.

There are whole cultures and ways of thinking built around production. The children of engineers who worked on xyz v1.0 have a genuine advantage when its time to work on xyz v2.0. There is a lot of tacit knowledge in these engineering fields and you have a huge advantage in knowledge retention if you can maintain unbroken chains of succession.

You can't achieve the top levels of ability (decades of experience, generational knowledge) if you are whip-sawing production to and fro across the globe every 10 years.

There are also cross pollination effects. Being in the same community with as many related fields as possible (co-located) is what drives cross-pollination and mobility of ideas and people between industries.

Think how many countries have tried to copy "silicon valley" and failed, and _why_ they failed.

What I'm saying is that technology is built by _people_ and there are human reasons why having local capacity is beneficial for all the related industries in the area.

xyzzy123··on CXMT has been offering DDR4 chips at about half the prevailing market rate
This is fine as long as the supply chain is, in fact, diversified.
xyzzy123··on CXMT has been offering DDR4 chips at about half the prevailing market rate
You become dependent on the supplier.

The downside in general is that other countries lose production capacity in steel, heavy industry, semiconductors, machine tools etc - industries that took decades to build and can't be easily replaced.

Also they gradually lose the ability to meaningfully innovate in those sectors because there's no grounding against production reality anymore.

This has geopolitical consequences further down the line.

xyzzy123··on AI is not a coworker, it's an exoskeleton
Sure, but do you want abundance of software, or scarcity?

The price of having "star trek computers" is that people who work with computers have to adapt to the changes. Seems worth it?

xyzzy123··on Claude Sonnet 4.6
Even if code gets cheaper, running your own versions of things comes with significant downsides.

Software exists as part of an ecosystem of related software, human communities, companies etc. Software benefits from network effects both at development time and at runtime.

With full custom software, you users / customers won't be experienced with it. AI won't automatically know all about it, or be able to diagnose errors without detailed inspection. You can't name drop it. You don't benefit from shared effort by the community / vendors. Support is more difficult.

We are also likely to see "the bar" for what constitutes good software raise over time.

All the big software companies are in a position to direct enormous token flows into their flagship products, and they have every incentive to get really good at scaling that.

xyzzy123··on AI is destroying open source, and it's not even good yet
Would be cool if you could donate to maintainer's favourite bot to get bugs fixed.

Usually, getting stuff fixed on main is better than being forced to maintain a private fork.

xyzzy123··on "I Was a Director at Amex When They Started Replacing Us with $30K Workers" [video]
I think all of your points are valid and I can't really see any part if your argument that isn't at least directionally correct. But then I'm left wondering:

Why is protectionism working for China?

xyzzy123··on OpenClaw (ClawdBot) joins OpenAI
There are enthusiasts and early adopters using this in small businesses already. In terms of "practical niche" the use case I've seen so far is "n8n but you create workflows by talking to it" aka business glue to automate idiosyncratic things.

I think HN being mostly quite technical under estimate the latent demand for ad-hoc business automation by people who know what they want to happen but aren't comfortable writing code.

You could look at it as a generic replacement for many types of AI SaaS harness. Previously if you wanted to reduce the workload of an office worker say reading work orders (that arrive in 50 different formats via email, sometimes as pdfs or behind portal links) and entering them into job control, you would need to write a custom agent harness or use a SaaS. Now you can sort of "mold" this thing like clay and get it to do the job. Instead of writing an API integration for the job control system you can just give it the openapi spec. Instead of writing your business logic in code, you can describe it in English. If you are technical, you can work with it to turn parts of the workflow into code to reduce token spend or make them more deterministic.

Naturally, it has all the disadvantages of home built automation (typically limited reproducibility, less secure, not generalised).

There's a lot of jank and risk but, hiring people can be pretty hit and miss in that regard also so for small businesses it's not as "out of distribution" as you might think.

Corporate is a different story.

xyzzy123··on Ask HN: My OpenClaw doesn't respond. Anybody met with the same problem?
Run Claude Code on the node and ask it to debug the issue...
xyzzy123··on I'm not worried about AI job loss
I'm not very familiar with the field on a practical basis.

What parts of the job require judgement that is resistant to automation? What percentage of customers need that?

If the hours an accountant spends on a customer go from 4 per month to 1, do you reckon they can sustainably charge the same?

xyzzy123··on Ask HN: We're building a saving app for European savers and need GTM advice
I guess the problem is that it seems like an opaque financial product. ETFs or crypto are not useful for "saving" but the risk profile is at least a known quantity.

The return you quote is a numeric range but there's no corresponding number or signal you provide for risk and we have to trust you. So, how can we trust you? Independent audits and deposit guarantees.

xyzzy123··on Europe's $24T Breakup with Visa and Mastercard Has Begun
> I'm not privy to device-wide bypasses of Play Integrity that ship with Trusted Execution Environment (which is pretty much all ARM based Androids), Secure Element, and/or Hardware Root of Trust, but I'd appreciate if you have some significant exploit writeups (on Pixels, preferably) for me to look at?

Hi, you don't have the break the control on the strongest device. You only have to break it on the weakest device that's not blacklisted.

The situation is getting better as you note, but in the past the problem was that a lot of customers have potatos and you get a lot of support calls when you lock them out.

> think Pacemaker / Insulin monitoring apps; government-issued IDs; financial instruments like credit cards; etc

I agree with you on the need for trustworthy computing. I mainly disagree on who should ultimately control the trust roots.

xyzzy123··on Europe's $24T Breakup with Visa and Mastercard Has Begun
If you work on mobile apps you will notice that full attestation is too slow to put in the login path. [This might be better than it used to be, now in 2026].

I don't think a good security engineer would rely on atty as "front line" anti brute force control since bypasses are not that rare. But yeah you might incorporate it into the flow. Just like captchas, rate limiting, fingerprints etc and all the other controls you need for web, anyway.

I know I'm quibbling. My concern is that future where banks can "trust the client" is a future of total big tech capture of computing platforms, and I know banks and government don't really care, but I do.

xyzzy123··on Europe's $24T Breakup with Visa and Mastercard Has Begun
An important security measure for who, though? The servers at the bank should "never trust the client" in case the attestation is bypassed or compromised, which is always a risk at scale.

If it's an important safety measure _for me_, shouldn't I get to decide whether I need it based on context?

I think it's fair for banks to apply different risk scores based on the signals they have available (including attestation state), but I also don't want the financial system, government & big tech platforms to have a hard veto on what devices I compute with.

xyzzy123··on Eight more months of agents
Org processes have not changed. Lots of the devs I know are enjoying the speedup on mundane work, consuming it as a temporary lifestyle surplus until everything else catches up.

You can't saw faster than the wood arrives. Also the layout of the whole job site is now wrong and the council approvals were the actual bottleneck to how many houses could be built in the first place... :/

xyzzy123··on Goldman Sachs taps Anthropic's Claude to automate accounting, compliance roles
It seems like the best way the humans could compete is to be a bit more flexible on interpretations than a fully compliant AI would be allowed to be. This applies all the way from small to very large businesses.

AI for this stuff is sort of a double edged sword; eventually it'll be forced to be legible in ways that you can't force a human mind to be. There are lots of domains where in theory AI would be great, but in practice everyone actually wants a certain amount of ambiguity and confusion to exist.

xyzzy123··on Claude Code for Infrastructure
Agreed on all points. Local loops are faster and safer wherever possible.

But particularly for devops / systems focused work, you lose too much "test fidelity" if you're not integrating against real services / cloud.

xyzzy123··on Claude Code for Infrastructure
Many places have "dev", "test" "prod"... but IMHO you need "sandpit" as well.

From an ops point of view as orgs get big enough, dev wraps around to being prod-like... in the sense that it has the property that there's going to be a lot of annoyed people whose time you're wasting if you break things.

You can take the approach of having more guard rails and controls to stop people breaking things but personally I prefer the "sandpit" approach, where you have accounts / environments where anything goes. Like, if anyone is allowed to complain it's broken, it's not sandpit anymore. That makes them an ok place to let agents loose for "whole system" work.

I see tools like this as a sort of alternative / workaround.

xyzzy123··on Surely the crash of the US economy has to be soon
A bottle of water in the desert and a bottle of water in your fridge don't have the same value.
xyzzy123··on Over 36,500 killed in Iran's deadliest massacre, documents reveal
It looks a LOT like a CIA front.

EDIT: Sorry... that is too strong... "state aligned influence media". Note that the headline might be true, or it might not, but that source is quite glowy.

xyzzy123··on How I estimate work
You mean… by making a forward-looking estimates of cost, time-to-value, return? (even if it's implicit, not documented, vibes-based?).

When devs refuse to estimate, it just pushes the estimating up the org chart. Execs still have to commit resources and do sequencing. They’ll just do it with less information.

xyzzy123··on How I estimate work
How would you decide between doing project (a) this quarter, or project (b)?

If you cannot (or refuse to) estimate cost or probability of success in a timebox you have no way to figure out ROI.

To rationally allocate money to something, someone has to do the estimate.

xyzzy123··on How I estimate work
Sorry, I edited the plumbing refence out of my comment because I saw a sibling post that made a similar point.

I agree there is less uncertainty in plumbing - but not none. My brother runs a plumbing company and they do lose money on jobs sometimes, even with considerable margin. Also when I've needed to get n quotes, the variation was usually considerable.

I think one big situational difference is that my brother is to some extent "on the hook" for quotes (variations / exclusions / assumptions aside) and the consequences are fairly direct.

Whereas as an employee giving an estimate to another department, hey you do your best but there are realistically zero consequences for being wrong. Like maybe there is some reputational cost? But either me or that manager is likely to be gone in a few years, and anyway, it's all the company's money...

xyzzy123··on How I estimate work
The top down push for timelines is because:

In Australia, an SDE + overhead costs say $1500 / work day, so 4 engineers for a month is about $100k. The money has to be allocated from budgets and planned for etc. Dev effort affects the financial viability and competitiveness of projects.

I feel like many employees have a kind of blind spot around this? Like for most other situations, money is a thing to be thought about and carefully accounted for, BUT in the specific case where it's their own days of effort, those don't feel like money.

Also, the software itself presumably has some impact or outcome and quite often dates can matter for that. Especially if there are external commitments.

xyzzy123··on Internet Archive's Storage
Not a pro data guy but someone running something like what you're talking about for many years. These days 200TiB is "normal storage server" territory, not anything exotic. You can just do the most boring thing and it will be fine. I'm just running 1, tho. The hard parts are having it be efficient, quiet and cheap which always feels like an impossible triangle.

Yeah, resilvers will take 24h if your pool is getting full but with RAIDZ2 it's not that scary.

I'm running TrueNAS scale. I used to just use Ubuntu (more flexible!) but over many years I had a some bad upgrades where kernel & zfs stopped being friends. My rack is pretty nearby so for me, a big 4U case with 120mm front fans was high priority, it has a good noise profile if you replace with Noctuas, you get a constant "whoosh" rather than a whine etc.

Running 8+2 with 24tb drives. I used to run with 20 slots full of old ex-cloud SAS drives but it's more heat / noise / power intensive. Also, you lose flexibility if you don't have free slots. So eventually ponied up for 24tb disks. It hurt my wallet but greatly reduced noise and power.

  Case: RM43-320-RS 4U

  CPU: Intel Xeon E3-1231 v3 @ 3.40GHz (4C/8T, 22nm, 80W TDP)
  RAM: 32GB DDR3 ECC
  Motherboard: Supermicro X10SL7-F (microATX, LGA1150 socket)
    - Onboard: Dual Intel I210 1GbE (unused)
    - Onboard: LSI SAS2308 8-port SAS2 controller (6Gbps, IT mode)
    - Onboard: Intel C220 chipset 6-port SATA controller

  Storage Controllers:
    - LSI SAS2308 (onboard) → Intel RES2SV240 backplane (SFF-8087 cables)
    - Intel C220 SATA (onboard) → boot SSD

  Backplane:
    - Intel RES2SV240 24-bay 2U/3U SAS2 Expander
    - 20× 3.5" hot-swap bays (10 populated, 10 empty)
    - Connects via Mini SAS HD SFF-8643 to Mini SAS SFF-8087 Cable, 0.8M x 5

  Boot/Cache:
    - Intel 120GB SSD SSDSC2CW120A3 (boot drive, SATA)
    - Intel Optane 280GB SSDPED1D280GA (ZFS SLOG device, NVMe)

  Network:
    - Intel 82599ES dual-port 10GbE SFP+ NIC (PCIe x8 add-in card)
It's a super old box but it does fine and will max 10Gbe for sequential and do 10k write iops / 1k random read iops without problems. Not great, not terrible. You don't really need the SLOG unless you plan to run VMs or databases off it.

I personally try to run with no more than 10 slots out of 20 used. This gives a bit of flexibility for expanding, auxiliary pools, etc etc. Often you find you need twice as much storage as you're planning on directly using. For upgrades, snapshots, transfers, ad-hoc stuff etc.

Re: dedup, I would personally look to dedup at the application layer rather than in the filesystem if I possibly could? If you are running custom archiving software then it's something you'd want to handle in the scope of that. Depends on the data obviously, but it's going to be more predictable, and you understand your data the best. I don't have zfs de-dup turned on but for a 200TiB pool with 128k blocks, the zfs DDT will want like 500GiB ram. Which is NOT cheap in 2026.

I also run a 7-node ceph cluster "for funsies". I love the flexibility of it... but I don't think ceph truly makes sense until you have multiple racks or you have hard 24/7 requirements.

← PreviousPage 5 of 34Next →