Hacker News Reader

RobinLon Libpostal: C library for parsing/normalizing street addresses around the world

There are many useful applications of libpostal, and it's an impressive library, but one I would caution against is for the purpose of address matching, at least as the 'primary' approach.

The problem is the hardest to parse addresses are also often the hardest to match, making the problem somewhat circular. I wrote about this more in a recent blog on address matching: https://www.robinlinacre.com/address_matching/

RobinLon Spending Too Much Money on a Coding Agent

This is why unlimited plans are always revoked eventually - a small fraction of users can be responsible for huge costs (Amazon's unlimited file backup service is another good example). Also whilst in general I don't think there's much to worry about with AI energy use, burning $24k of tokens must surely be responsible for a pretty large amount of energy

RobinLon Ask HN: What's the last non-obvious skill that made you better at your job?

Writing a blog. Many of my posts are things I've learnt at work, or arguments I've failed to make it meetings. By writing it down, I can pin down the argument better and share my thoughts in advance.

RobinLon Building Accurate Address Matching Systems

Thanks - appreciate the kind words!

RobinLon Ask HN: What Are You Working On? (June 2025)

I'm working on a free high performance address matching (geocoding) library. As it turns out I blogged about it just today: https://www.robinlinacre.com/address_matching/

RobinLon Model Once, Represent Everywhere: UDA (Unified Data Architecture) at Netflix

> "the video game people model it differently because they're doing different analysis, and if you unify the base layer to support everybody's type of analysis, it's not going to change that there's still a bunch of independent, not-talking-to-each-other analysis going on"

This resonates. Moreover, it's very easy for architects to assume that because different areas of the business use data about the 'same' thing, the thing must be the same.

But often the analysis requires a slightly different thing. Like: we want a master list of prisons. But is a prison a building, a collection of prisoners (such that the male prison and the female prison on the same site are different prisons), or the institution with that name managed under a particular contract?

RobinLon How Anthropic teams use Claude Code

Codex also had this feeling to me, it's like a control centre for a repo, but it's easy to imagine that being the whole computer.

There's a kind of inversion in how you do things: e.g. rather than navigating to settings until you find the screen time controls and changing the settings, you explain in natural language what to do, the agent asynchronously proposes the solution, and you just click accept or reject.

RobinLon Trying to teach in the age of the AI homework machine

When hiring, I would very much like to hire people who have figured out how to learn things for themselves using whatever techniques work for them, and don't need nannying.

So I'm perfectly happy with a system of higher education that strongly rewards this behaviour

RobinLon Ask HN: What are you working on? (May 2025)

I've been experimenting with whether I can use LLMs to produce interactive maths explainers for kids. There are a few examples here: https://rupertlinacre.com/

Unless I'm missing something, it's amazing how few free, _high quality_ materials are online.

Ultimately I'm interested in two things: genuinely fun games that make you do some maths, and quality visualisations that help make concepts easier to learn

RobinLon Void: Open-source Cursor alternative

I agree that on the face of it this is extremely useful. I tried using it for multiple libraries and it was a complete failure though, it failed to crawl fairly standard mkdocs and sphynx sites. I guess it's better for the 'built in' ones that they've pre-indexed

RobinLon Cline is quietly eating Cursor's lunch and changing how we vibe code

Is the key here that with Cline, it's passing huge amounts of the code into context, unlike Cursor? So in some sense it's not surprising it does a better job, because it's executing much more expensive API calls to the LLM provider

RobinLon OpenAI reaches agreement to buy Windsurf for $3B

How long ago was that? 6 months ago I switched from VS Code to Cursor, which at the time was FAR superior to Copilot. Around a month ago I switched back to VS Code, and found there's not much difference any more. Autocomplete in VS Code is still less good, but the agent mode in VS Code feels pretty similar to Cursor's (albeit a little slower, perhaps).

To be honest I think both are quite limited by context length (in that they try to limit the context they send to the LLM and hence cost), and so I find myself using Gemini 2.5 in AI studio with the 1m context length, and asking it to generate instructions for Copilot (which seems to work pretty well)

RobinLon Suno v4.5

https://x.com/sama/status/1899535387435086115?lang=en-GB

It's personal taste, but this is significantly better than the last couple of fiction books I've read (which were both well reviewed).

I think it's good enough that it's hard to argue is emotionally vacuous, unless you define that to mean 'it was written by a machine'

I think increasing we'll find AIs are extremely good at emotional 'manipulation' (which I mean in the same sense has how a good tearjerker is in some sense emotionally manipulative).

RobinLon A cheat sheet for why using ChatGPT is not bad for the environment

For context thats's about equivalent to 100 transatlantic flights

RobinLon Vibe Coding: Developer Slot Machines (Cursor, Windsurf)

Yes, you can. It's a relatively recent addition. The original commenter probably doesn't see much difference between copilot and Cursor because copilot has caught up quite a lot in the last couple months

RobinLon Instant SQL for results as you type in DuckDB UI

Can you go the other way? (E.g. edit the above and turn it back into SQL string)

I've used sqlglot to do this in the past, but doing it natively would be nice

RobinLon Instant SQL for results as you type in DuckDB UI

Agree, definitely amazing feature. In the Python API you can get somewhere close with this kind of thing:

input_data = duckdb.sql("SELECT * FROM read_parquet('...')")

step_1 = duckdb.sql("SELECT ... FROM input_data JOIN ...")

step_2 = duckdb.sql("SELECT ... FROM step_1")

final = duckdb.sql("SELECT ... FROM step_2;")

RobinLon Geocoding APIs compared: Pricing, free tiers and terms of use

My experience is also that (in the uk at least) it can't be done accurately with openstreetmap data.

For context, I've tried it! I've been working on a free library for geocoding UK addresses quickly and accurately. It comes with the caveat that you need access to the dataset of all addresses you're geocoding against - which could be your own list, or a commercial product like addressbase: https://github.com/robinL/uk_address_matcher/

RobinLon Don't force your kids to do math

There is a game called dragonbox algebra which I'm currently working through with my son and is an absolutely fantastic approach to this problem. Sadly its now part of a horrendous subscription service and is hard to access. I find it really sad that we've had computers for decades and there are so few good maths games like this.

RobinLon Waiting 100 years for a home isn't a housing crisis, it's a moral collapse

Preface: I'm a yimby who wants to see far more building, especially building upwards in London, but also more general building elsewhere.

That said, a 100 year wait for social housing in Westminster is not surprising, and not as bad as it sounds. A typical professional couple in London could not afford a family home in Westminster, and it's extremely common for people in their 30s to move out of the centre when they have kids. I don't think this should be any different for social tenants: they should certainly not have to wait 100 years, but it seems reasonable for them to be offered housing outside of Westminster. The Guardian article I read about the 100 year wait specifically mentioned 3 and 4 bed family homes, and my immediate reaction was that no one else can get them either!

I think the bigger problem is the lack of affordable housing particularly for young families (social or otherwise) outside of London

RobinLon Gemini 2.5 Flash

Does anyone know how this pricing works? Supposing I have a classification prompt where I need the response to be a binary yes/no. I need one token of output, but reasoning will obviously add far more than 6 additional tokens. Is it still a 6x price multiplier? That doesn't seem to make sense, but not does paying 6x more for every token including reasoning ones

RobinLon AI used for skin cancer checks at London hospital

The referral is just because a non expert can't be sure. The cost of the referral is relatively small, but the cost of getting it wrong is large.

Someone I know recently had a referral - it's pretty light touch, you just get a prompt appointment, and they do a minor op to remove the mole, and send it to the lab for testing. Luckily in their case, it wasn't cancer. But nothing in the process seemed weird, it was just the way of the GP escalating it because they couldn't be sure. Hypothetically, if the AI had been able to diagnose with higher certainty than the GP, all of this could have been avoided, so definitely room for improvement.

RobinLon Apache ECharts

As a big user of vega lite I think that's fair. I think it really shines when used by data vis experts, where charts need to be precise, such as in research and analysis contexts. For something like a simple a metrics dashboard I think I'd agree that it may be difficult for devs.

RobinLon GitHub Copilot Pro+

I've had problems with pylance and auto complete on Cursor for months for (what I think is) the same reason - so I'm not sure the crackdown has only started in the last couple days

RobinLon GitHub Copilot: Agent mode and MCP support rolling out to all VS Code users

Interested in experiences switching back to VS Code from Cursor. I've recently (past couple weeks) been trying out copilot again on Vs code insiders, which has had agent mode for a while.

My perception is that it's caught up a lot. It still seems a bit slower to run agent, and slower to apply diffs. As the auto complete is a little worse. But it's no longer a no brainer to spend the extra on the cursor subscription.

For harder problems, I've also found myself reaching for long-context Gemini 2.5 pro using files-to-prompt, meaning I've been relying on copilot/cursor for less complex tasks

RobinLon Why does Britain feel so poor?

To make matters worse, there's also litte evidence that the increases in spending on SEND provision have led to better outcomes.

I can't help feeling like this is a vicious cycle - the lack of community facilities is causing greater isolation, causing a rise in health needs and so on.

RobinLon Ask HN: What are you working on? (March 2025)

- A FOSS Python package for record linkage and deduplication at scale called Splink: https://moj-analytical-services.github.io/splink/

- A library for fast, accurate matching (geocoding) of UK addresses that uses Splink under the hood: https://github.com/RobinL/uk_address_matcher

- An npm library that generates maths mental arithmetic problems that align to the UK national curriculum, that can be used to power maths games: https://github.com/RobinL/maths-game-problem-generator

- A breakout maths game that uses the above: https://github.com/RobinL/maths-game-problem-generator

RobinLon The Mediocrity of Modern Google

Agree, they write it off as:

> Gemini 2.5 barely made a ripple in the broader tech conversation.

But my sense (vibes from using it) is it's the very best model right now (though I have not tried the $200 a month exclusive openai offerings)

RobinLon 4o Image Generation

It's very impressive. It feels like the text is a bit of a hack where they're somehow rendering the text separately and interpolating it into the image. Not always, I got it to render calligraphy with flourishes, but only for a handful of words.

For example, I asked it to render a few lines of text on a medieval scroll, and it basically looked like a picture of a gothic font written onto a background image of a scroll