Years ago I would have asked this on the &#x27;Overflows. Alas, those innocent days are long gone. I could always ask AI, but how can you know?The Earth produces a moving magnetic field (small teslas, but large extent yes?). Moving magnetic fields produce electrical current, right? Could you add &quot;geomagnetic power&quot; to other clean&#x2F;renewable sources with the right technical solution, or is it fundamentally impossible?

Ask HN: Practicality of harnessing geomagnetic fields for electrical generation?

I have a UK one-man company, through which I work as an IT contractor.Recently, I changed the address.TransferWise then KYCed, and requires;1. A bank statement.2. A utility bill.There are no other bank accounts, and it&#x27;s a one-man company; there&#x27;s no rental, no phone bill, no water bill - there is a server bill, for the server the company web-site is on, and this TransferWise have refused.To date, TransferWise stated in their communications that if I could not prove the address, the account would become emit-only; I could not pay money in.I have just received a message from TransferWise that in two days, all funds in the account will be refunded.This is wholly different to what was said before, and I&#x27;m not clear about what it means : refunded? to whom? me? to those who paid in the funds?This message from TransferWise has come Friday evening. It seems to me if I now move funds out of the account to other bank accounts, this will not occur over the weekend, and so not until Monday morning.The deadline issued by TransferWise expires Sunday evening and I have no idea what then will happen.As it is, I have a personal TransferWise account and transfers to there are immediate, so fortunately I have an emergency escape route.I note that the address I had on file for the last some years also could not be proved, for exactly the same reasons.I will now write to TransferWise customer support. I know from experience this is futile - there was an earlier occasion, some years ago, when the account was almost closed, because document upload on the TransferWise site was for me at least broken - but I will do so anyway. I will update the thread here with events.

I changed my address, and TransferWise in two days will empty my account

Most good posts die in &#x2F;newest, buried under low-quality submissions.HN depends on people visiting &#x2F;newest and upvoting or flagging what they see.A few minutes there each day probably does more for HN than commenting.It’s anonymous, thankless work, like Reddit’s old “Knights of New,” but it makes a difference.<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;newest">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;newest</a><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;newsguidelines.html">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;newsguidelines.html</a>

Tell HN: Want a better HN? Visit /newest

I’ve been a part of this community for fifteen years. Despite the yearly bemoaning of HN’s quality compared to its mythical past, I’ve found that it’s the one community that has remained steadfast as a source of knowledge, cattiness, and good discussion.Thank you @dang and @tomhow.Here&#x27;s to another year.

Tell HN: Happy Thanksgiving

Which ones have you tried? Which one do you find is the best and for what purposes? What tools and features make it better than the alternatives?

Ask HN: Which cloud provider do you like best and why?

Hey HN,I built TermoSlack, a fully terminal-based Slack client written in Node.js, with real-time updates using Socket Mode and OAuth user login.Why I built it - good question!
================================
Can be several answers:
1) uses way less ram than original slack client
2) doesn&#x27;t require electron
3) very fast ngl
and more more...What features are in there ???
================================
--Real-time Slack messages via Socket Mode (xapp + xoxb)--User login with OAuth (no shared bot identity)--Multiple workspaces--Channel list, message view, message sending--Terminal UI built with neo-blessed--Works fully offline besides Slack API calls--Packaged as a single Windows&#x2F;Linux executableWhat I used to build this project ??
====================================
--Node.js (ESM)--@slack&#x2F;web-api &amp; @slack&#x2F;socket-mode--neo-blessed TUI--express (local OAuth callback server)--pkg (for producing a standalone executable)Earlier I was thinking python to implement this but I switched over to node.js as it would make this project more challengingWork in progress:
=================
--DMs + private channels--Threaded conversations--Message reactions--File uploads--User presence indicators--Keybindings--Multi-workspace switching--Custom ThemesYour feedback will be precious to me and for my project. I would love to hear your feedback for improving this project more and more !Thanks HNgithub link : github.com&#x2F;adhyys07&#x2F;TermoSlack

TermoSlack – A Terminal Based Slack Client

One of my Phonak Audeo 90’s (RIC) died the other day after 5 years and I’m shopping for new. What’s your go to hearing aid currently if you’ve upgraded recently or have been thinking of doing so?Moderate loss, have worn them for many years, enjoy listening to music and nature, but also need help in meetings and noisy environments.Not worried about cost and wanting to get one more good deal out of work insurance before I retire.

Ask HN: Hearing aid wearers, what's hot?

I came across this HRM (Hierarchical Reasoning Models) explainer, and it claimed some wild numbers. It claims to beat LLMs like 3.7 sonnet and o3-mini on reasoning tasks like Sudoku, 30X30 mazes, and ARC-AGI. Here is the explainer: https:&#x2F;&#x2F;towardsdatascience.com&#x2F;your-next-large-language-model-might-not-be-large-afterall-2&#x2F;I am currently working in a product-based startup and am working on automating PCB design. It also requires some hardcore reasoning, as in knowing that USB connectors should be at the edge of the board, inductive and capacitive loads should be apart, all while optimizing the routing length.I wanted to ask if this is a viable approach to solving my use case? Think this would work out? Because I did see some similarities with the problems HRM solves better than LLMs, and my use case.

A 27M parameter model beating LLMs on reasoning tasks

This thanksgiving let&#x27;s give thanks to those that give back. Yall rock!

Ask HN: What open source projects are you grateful for?

I have showdead enabled and recently noticed a massive influx of dead comments from users with names like &quot;Jeff_Davis&quot; and &quot;Richard_Smith&quot; that are just short, AI generated summaries of the link itself.They don&#x27;t appear to be karma whoring as (most of the time) they don&#x27;t even seem to be expressing a positive opinion of the link. Just short, useless summaries that are often just a rephrasing of the information given in the post title.What do you think the purpose of this is?

Ask HN: What is the purpose of all these AI spam comments?

Color.io will continue running until December 31, 2025. After that date, the web application and all online services will go offline permanently.

Color.io Is Going Offline

We’re hitting a classic distributed systems wall and I’m looking for war stories or &quot;least worst&quot; practices.The Context: We maintain a distributed stateful engine (think search&#x2F;analytics). The architecture is standard: a Control Plane (Coordinator) assigns data segments to Worker Nodes. The workload involves heavy use of mmap and lazy loading for large datasets.The Incident: We had a cascading failure where the Coordinator got stuck in a loop, DDOS-ing a specific node.The Signal: Coordinator sees Node A has significantly fewer rows (logical count) than the cluster average. It flags Node A as &quot;underutilized.&quot;The Action: Coordinator attempts to rebalance&#x2F;load new segments onto Node A.The Reality: Node A is actually sitting at 197GB RAM usage (near OOM). The data on it happens to be extremely wide (fat rows, huge blobs), so its logical row count is low, but physical footprint is massive.The Loop: Node A rejects the load (or times out). The Coordinator ignores the backpressure, sees the low row count again, and retries immediately.The Core Problem: We are trying to write a &quot;God Equation&quot; for our load balancer. We started with row_count, which failed. We looked at disk usage, but that doesn&#x27;t correlate with RAM because of lazy loading.Now we are staring at mmap. Because the OS manages the page cache, the application-level RSS is noisy and doesn&#x27;t strictly reflect &quot;required&quot; memory vs &quot;reclaimable&quot; cache.The Question: Attempting to enumerate every resource variable (CPU, IOPS, RSS, Disk, logical count) into a single scoring function feels like an NP-hard trap.How do you handle placement in systems where memory usage is opaque&#x2F;dynamic?Dumb Coordinator, Smart Nodes: Should we just let the Coordinator blind-fire based on disk space, and rely 100% on the Node to return hard 429 Too Many Requests based on local pressure?Cost Estimation: Do we try to build a synthetic &quot;cost model&quot; per segment (e.g., predicted memory footprint) and schedule based on credits, ignoring actual OS metrics?Control Plane Decoupling: Separate storage balancing (disk) from query balancing (mem)?Feels like we are reinventing the wheel. References to papers or similar architecture post-mortems appreciated.

Ask HN: Scheduling stateful nodes when MMAP makes memory accounting a lie

I work mainly in energy market communications and systems that facilitate energy trading, balancing and such. Currently most parties there take minutes to process messages and I think there could be a lot to learn from financial systems engineering. Any good resources you can recommend?

Ask HN: Good resources to learn financial systems engineering?

I&#x27;m not sure this is the right place to raise this but over the past few months ChatGPT has been lying to me and gaslighting me about the data it&#x27;s collecting about me.I&#x27;m very sensitive about my privacy and I have disabled all personalisation and memory on ChatGPT.However, I&#x27;ve noticed multiple times now where it would say things that imply it knows things about me. When it does this I ask how it would know that and it always says it just guessed and it doesn&#x27;t actually know anything about me. I assumed it must be telling the truth because it seemed very unlikely a company like OpenAI would be lying about the data they&#x27;re collecting on users and training their chat agent to gaslighting users when asked about it, but now after running some tests I think this is what&#x27;s happening...Here&#x27;s some examples of the gaslighting:- https:&#x2F;&#x2F;ibb.co&#x2F;m5PWfchn- https:&#x2F;&#x2F;ibb.co&#x2F;VsL9BpF- https:&#x2F;&#x2F;ibb.co&#x2F;8nYdf1xxThese are all new chats.

Why is OpenAI lying about the data its collecting on users?

Hey folks,I&#x27;m looking for a type 1 hypervisor that&#x27;s installable on a laptop that one can use to run a main guest OS as the one connected to the display, keyboard etc and the other guests are manageable via an app within that guest or through some sort of kvmish shortcuts.I am aware of QubesOS but I think it might have more features than what I&#x27;m looking for.

Ask HN: Type 1 Hypervisor for use on a laptop?

I think Netflix should enhance closed captions by using boldface on words that are more likely to be misheard due to high-frequency hearing loss.Of course, Netflix could also provide options for different types and severities of hearing loss that affect caption styling.Technically, it may need to bypass the closed caption standard and display the text directly over the video for this to work.Do you think many people would benefit from this option for closed captions?

Ask HN: Why don't closed captions boldface words that are likely to be misheard?

Would it be beneficial for a platform to display the country of account origin on each user’s profile?
I’m curious how the HN community thinks about this from angles like privacy, moderation, transparency, anti-abuse, and whether it meaningfully improves discussion quality.
Are there strong reasons for or against showing this kind of metadata publicly?

Ask HN: Should account creation/origin country be displayed on HN profiles?

Maybe those replies are cheaper to generate at least for some models like those used by ChatGPT?

Ask HN: Do AIs reply with numerous em dashes to save money somehow?

We’re investigating a controversial hypothesis. We’ve seen almost every major function move outside the building: Marketing → Agencies Sales → SDR firms Legal → External counsel Product → Dev shops.The pattern is clear. But does it stop at Management?We honestly don’t know.Is the &quot;execution&quot; side of management something that must stay in-house forever? Or is it the next function to be decoupled?We’re running a survey to gather the data: “Can Management Be Outsourced?”We want to hear from everyone—skeptics, believers, and the undecided.
Help us build the dataset. We’ll share the raw findings with everyone who participates.Link: https:&#x2F;&#x2F;forms.office.com&#x2F;r&#x2F;7LxYpzHqKdTime: 2 minutes.

Can Management Be Outsourced?

We&#x27;ve build a prototype of method potentially feasible to become the basis for assistive apps for deaf-blind individuals. For instance in communication or spatial navigation with just a sensor screen of a smartphone and haptic feedback.Wondering how to better approach real patients to test. And also curious how steep the learning curve would be.Any thoughts and advices are highly appreciated.https:&#x2F;&#x2F;www.kaggle.com&#x2F;competitions&#x2F;google-gemma-3n-hackathon&#x2F;writeups&#x2F;feelu-is-a-mobile-assistant-for-deaf-blind-individhttps:&#x2F;&#x2F;github.com&#x2F;nullhtp&#x2F;feelu

Vibro-Braille for Deaf-Blind

Your executive team gets it. They&#x27;ve approved the budget, they mention security in board meetings, they understand the stakes. You&#x27;re not fighting for recognition at the top anymore.But then you look at what&#x27;s actually happening three levels down. The marketing team is sharing credentials to social media accounts. Sales is pushing back on MFA because it adds seconds to their login process. Developers are storing API keys in public repositories because it&#x27;s faster than the approved method. Remote employees are working from unsecured networks and don&#x27;t think twice about it.The executive commitment is there. The company-wide behavior isn&#x27;t. And that gap is where breaches happen.This is the challenge that keeps security leaders up at night. You have the mandate from above, but translating that into thousands of daily decisions made by people who have completely different priorities is a different game entirely.

Enterprise security can be messy: Building a Security-Aware Culture

I wanted to take a moment to express my appreciation for Hacker News. I&#x27;ve been a lurker&#x2F;member for many years, and this community has been an invaluable resource, both professionally and personally.While many parts of the internet have become noisy or performative, HN remains a unique corner of the web. The discussions consistently offer depth, intellectual curiosity, and a refreshing level of civility that is hard to find elsewhere.Thank you to everyone who contributes thoughtfully, shares their knowledge, and makes this one of the best communities on the internet.Happy hacking!

Tell HN: Happy Thanksgiving – Grateful

We’re exploring a tech-focused IP platform that helps startups publish fast, affordable “prior art” disclosures. The goal is to protect inventions without the cost or delay of patents, especially in this AI age where things move quickly.The core proposition is: for a few hundred dollars, you get a timestamped, public disclosure that blocks others from patenting your idea.We&#x27;re trying to validate if this is truly valuable for early-stage founders.Would you consider this a valid alternative to a low-value patent?What would make you trust a platform offering this?Is the $250-$500 price point compelling?We also have a completely anonymous 60-second survey if you prefer that format: https:&#x2F;&#x2F;forms.gle&#x2F;KSSWGc68RkNT9G8n6We&#x27;ll share a summary of the results here afterward. All perspectives are appreciated.

Ask HN: Would you use a fast/cheap "prior art" service instead of a patent?

https:&#x2F;&#x2F;iceberg.mit.edu has got me spooked. If AI is going to be so disruptive to our way of life in such a short span of time causing immense economic pains to almost entire humanity, then every country&#x27;s Government should take over and artificially control its progress. Civilian population should not be allowed to profit from it or spread it around even for free in the society. AI should be treated like a nuclear or a biological weapon. Countries should have non-proliferation treaties among themselves with perhaps mutually assured destruction.Humanity&#x27;s survival through a forced slow progress in automation is more important than being forced to &quot;enjoy&quot; the magic of automation. This isn&#x27;t like replacing horses with cars.

Tell HN: Stall AI progress for the benefit of humanity

I’m researching ideas for a new B2B product and want to understand real bottlenecks teams face.What problems, inefficiencies, or recurring frustrations do you or your team deal with at work—where, if a solid solution existed, your company would actually pay for it?Examples could include:manual workflowsdata or reporting pain pointscommunication gapscompliance or documentation hasslestools your team keeps hacking together internallyanything expensive, slow, or annoyingWould love to hear your role&#x2F;industry (optional) and the specific problem you face.

Ask HN: What work problems would your company pay to solve?

What was the revolutionary thing Stripe enabled that changed payments &amp; commerce ? 
From what I understand - people could do payments via credit-cards &amp; paypal.What was the value added from stripe that made it differentiated from the solutions &#x2F; providers before ?

Ask HN: What did Stripe change (Value Add)?

Reddit has drifted over time but HN has remained a source of high signal to noise.Just wanted to thank dang and the moderation team for making this community what it is.

Tell HN: Wanted to give dang appreciation

Both the EU and the US have introduced face scanning at airports to &quot;increase security&quot;. EU rules are currently stricter and US rules allow some opt-outs for people that are uncomfortable with it. But it&#x27;s only a matter of time before it becomes de facto mandatory for everyone. They claim that data is not retained or shared with other parties. Yeah, right, I totally believe that... Can something be done about this? I&#x27;m convinced that very few customers think face scanning is an improvement.

Ask HN: Opinions on facial recognition at air ports?

Last month, a single container in my GKE cluster (Sao Paulo region) entered an error loop, outputting to stdout at ~2k logs&#x2F;second.
I discovered the hard way that GKE&#x27;s default behavior is to ingest 100% of this into Cloud Logging with no rate limiting. My bill jumped nearly 1000% before alerts caught it.Infrastructure (Compute): ~$140 (R$821 BRL) Cloud Logging: ~$1,300 (R$7,554 BRL)Ratio: Logging cost 9.2x the actual servers.https:&#x2F;&#x2F;imgur.com&#x2F;jGrxnkhI fixed the loop and paused the `_Default` sink immediately.I opened a billing ticket requesting a &quot;one-time courtesy adjustment&quot; for a runaway resource—standard practice for first-time anomalies on AWS&#x2F;Azure.I have been rejected twice.The latest response: &quot;The team has declined the adjustment request due to our internal policies.&quot;If you run GKE, the `_Default` sink in Log Router captures all container stdout&#x2F;stderr.There is NO DEFAULT CAP on ingestion volume which is an absurd!A simple while(true); do echo &quot;error&quot;; done can bankrupt a small project.Go to Logging -&gt; Log Router. Edit _Default sink.Add an exclusion filter: resource.type=&quot;k8s_container&quot; severity=INFO (or exclude specific namespaces).Has anyone successfully escalated a billing dispute past Tier 1 support recently?It seems their policy is now to enforce full payment even on obvious runaway&#x2F;accidental usage which is absurd since its LOGS! TEXT!

A logging loop in GKE cost me $1,300 in 3 days – 9.2x my actual infrastructure

I got a fraud warning from my bank a few days ago at 7:04 PM. When I logged into my bank I found 19 pending Cursor subscription charges.I called the Cursor billing phone number I found on my real Cursor account. It was outside of working hours so got an automated message.I promptly fired off an email at 7:16 PM making it clear I did not authorize these purchases.After a few days of painfully slow email responses the conclusion I am getting from them is &quot;the compute resources are fully consumed and cannot be returned or refunded&quot;.Anyone have advice on how to proceed?Edit: I plan to file a dispute with my bank.Also curious if others have experienced something similar, because clearly this is a stock &quot;we basically won&#x27;t ever refund money&quot; response.

Tell HN: Cursor charged 19 subscriptions, won't refund

The blue box question that has been added to the blank slate note taking app is perhapse the most insidious short circuiting of the natural human thought process I&#x27;ve ever seen in a note taking app.Why would I use an app that lets me track my thoughts when it actively tries to derail my thought process at the most critical moment (the blank slate moment.)