HNHacker News
TopNewBestAskShowJobs

lhnz

5,754 karma · joined June 16, 2010

www: https://sebinsua.com

github: https://github.com/sebinsua

gist: https://gist.github.com/sebinsua

twitter: https://twitter.com/sebinsua

[ my public key: https://keybase.io/sebinsua; my proof: https://keybase.io/sebinsua/sigs/h5SINjJKRvoKp0ggBhJuVn61d2Zfb3sbIp7ChK3OlTY ]

submissionscomments
lhnz··on EU opens proceedings against X over efforts to combat information manipulation
I think it's great that we don't have Chinese-style censorship of the internet but disagree that "there really isn't much to talk about (yet)". I think there is good reason to believe that many governments around the world are interested in making laws to control the dissemination of various kinds of speech, and whether or not we have the full information on individual cases this shouldn't stop us from commenting on our overall rights around speech and what we believe these should be. Personally, I don't wish to self-censor over this, as I think it's important.

I live in a country with even worse speech laws than those in the US and I wish people stood up for the kinds of speech that should be protected instead of waiting until our lawmakers pass draconian laws and then complaining.

lhnz··on EU opens proceedings against X over efforts to combat information manipulation
The press release itself says that the formal proceeding concern "the dissemination of illegal content in the context of Hamas' terrorist attacks against Israel" and the person you refer to as "some guy" is a adjunct professor of law at a few institutions, as well as a legal fellow of the Adam Smith Institute where his research interests have focused on free speech and related technology regulation.
lhnz··on EU opens proceedings against X over efforts to combat information manipulation
Even though we don't know everything about what is happening in this particular instance, I do think there should be a global conversation on the ability for state-level actors to directly or indirectly pull strings in order to coerce Twitter into censoring information on their behalf.

I think this should be of particular importance to the US which has fudamental rights protecting freedom of expression.

I'm very worried about co-ordinated lawfare against Musk-operated companies being used as a tool to force his compliance with systems that exist to get around laws against the US government's direct involvement in censorship. I'm also worried about the EU's lurch towards Chinese-style censorship of social media. And, I think we should be able to talk about it now while we still can.

lhnz··on EU opens proceedings against X over efforts to combat information manipulation
Apparently the dissemination of illegal content refers to the October 7th videos of the Hamas terrorist attack against Israel that galvanized American support for Israel [0].

I personally don't think it's in people's interest to have this kind of thing hidden from them and I really dislike the EU for attempting to censor this.

Musk is really under a lot of financial pressure and it might be a losing battle to try to fight the EU over this in court. To be honest, I hope he is able to pull out from the EU without cratering the company, but he is now being pulled into lawfare with multiple actors across the globe so Twitter might have no choice but to accept is role as a conduit of censorship for state-level actors.

  [0] https://twitter.com/prestonjbyrne/status/1736707341070860689
lhnz··on Building Node.js applications without dependencies
Yes, good shout -- I do this sometimes, too. The only thing is that it doesn't work if what you're releasing is a shared package or installable and is not part of an application.
lhnz··on Building Node.js applications without dependencies
I have a question for you all:

When you're building something and you notice one of your dependencies has a bug or is missing a key feature that you need, do you (a) PR the fix into the dependency and then try to "harrass" the maintainer into merging it for you, (b) publish your own fork of the dependency with the necessary fix, (c) inline the source code for the dependency into your project, effectively taking it on as if it's your own code, (d) completely rewrite the dependency either as a separate package you control or built directly into your own project, or (e) code around the problem / do a hack?

I find that often maintainers are so over-worked that it's practically impossible to get a merge in a timely manner, and this leads me to rely on a fork until the PR eventually gets merged. However, I think creating a new package under either your own ownership or the company you work for is often really bad as it can become a kind of hidden technical debt. Nowadays, I definitely consider inlining as a way to capture ownership of the technical debt in a way that is highly visible, but this can add 100s or 1000s of lines of code to a project and if eventually the upstream project moves on you don't get the benefits of their changes without removing the inlined code and untangling any changes that were made to it.

The only other approach I've seen is the 'hack' approach, in which you try to dodge the bug or semantic issue. Honestly, that might be the right thing to do in some situations, but it isn't very hygienic within a long-term project (unless you carefully maintain a TODO list of things that need 'correct' solutions).

lhnz··on Kagi Search – Paywalled articles indicator and improved weather widget
I'm also a happy Kagi user.

There are only two reasons I still visit Google:

  1. I'm based in the UK and often it shows me shops based
     in the US high up within my search results.
  2. Sometimes I specifically like to visit Google Shopping
     to find the lowest prices online for an item I'm trying
     to buy.
Once Kagi has improved these two areas I'll have no reason at all to ever visit Google (other than Google Maps or Google Translate).
lhnz··on Code is run more than read
The article is good on the whole but the title "Code is run more than read" seems false in many situations I've been in. A lot of software doesn't have users. Often the target users don't like the product enough and are just trialing it for you as a favour.

In these cases, it's not very meaningful to say it's being "run more than read", because either you should have made the code readable enough for other developers to delete it, modify it or rewrite it until users want to use it, or the code has to be valueless and able to be completely thrown away (the knowledge of how to build it is either written down elsewhere or tacit).

I also agree with what somebody else said about the times in which the features/fixes you create aren't seen by users but just benefit a business-to-business sales process. In these cases, you behave differently. You do the work because there's a very clear signal to your company meeting their expectations promptly.

I guess one way I would agree with the article is that once something has users/businesses and is somewhat readable you apply your extra effort on polishing the user experience as prioritising the happiness of your core users is generally a good idea.

lhnz··on AI and the Rise of Mediocrity
Hm, my own intuition is that it'll raise the baseline experience but then cause hyper-competition by human creators to exploit its powers to the limits while infusing further human creativity on top of what is mechanistically produced.

I do not think this will result in overall mediocrity. It'll lift those with no access to art or way to pay for human creativity away from absolute poverty, while allowing those that already had access to these things a more competitive market.

lhnz··on Freetar – an alternative front end for ultimate-guitar.com
Yes, it's very sad to see work you provided for free to a community used to exploit these people.
lhnz··on Before OpenAI, Sam Altman was fired from Y Combinator by his mentor
I am also English. :)
lhnz··on Before OpenAI, Sam Altman was fired from Y Combinator by his mentor
Well, I'm also English and I didn't read them that way. However, I do think that Paul is telling people that competing with Sam in certain domains would be extraordinarily difficult.

The other thing is that if you take a look at Paul Graham's blog posts, he used to regularly thank Sam at the bottom of these -- this isn't something you do if you don't like or respect someone. However, on the other hand, perhaps they fell out at some point? I can't personally make out that signal from the little data there is.

lhnz··on Before OpenAI, Sam Altman was fired from Y Combinator by his mentor
Really? It seems like a glowing appraisal. He seems to think that Sam is devestatingly effective at what he does.
lhnz··on Before OpenAI, Sam Altman was fired from Y Combinator by his mentor
If Paul Graham fired Sam Altman from YCombinator it's interesting that he appears to have such a favourable opinion of him [0].

However, personally, what I've taken away from this is that he is a much better strategic/tactical operator than many other high-flying executives and very capable of winning the respect and trust of a lot of smart people. I wouldn't expect OpenAI to be run by anybody that wasn't revered in this way; a lot of CEOs aren't saints.

[0] https://twitter.com/search?q=from:paulg%20since:2019-01-01%2...

lhnz··on A Node, TypeScript, TS-Node and ESM experience that works
I'm using `esyes` [0] by Yehuda Katz and the logic is simple and it works very well.

[0] https://github.com/starbeamjs/esyes

lhnz··on Ask HN: How to get back to programming Python?
I felt the same way and recently did the whole 92 questions of CodeSignal's Python Arcade section (https://app.codesignal.com/arcade/python-arcade). This made me completely fluent on modern Python syntax and the questions themselves are very easy so can be done quickly. You are really only re-learning Python's syntax as opposed to doing hard algorithmic problems or anything like that (although I do recommend doing this as you'll be able to touch on Python's standard library more).

I guess the one thing that you don't get to do with that approach is build something interesting or use `async-await` but it gets you fluent with the syntax again which is an important first step.

lhnz··on Children working as riders for food delivery apps – BBC investigation
Why don't they just allow children to work legitimately through the application, but add some safe-guarding controls and potentially notification to or authorisation by parents? This seems like a much better idea than to have a level of uncontrollable fraud that allows it through the backdoor and enriches rent-seekers.
lhnz··on Don't build AI products the way everyone else is doing it
I don't agree. You should be collecting data, running experiments and A/B tests and finetuning your own models, but you shouldn't be building CRUD apps to automate this yourself, even if you have two decades of experience doing so. You shouldn't be building out this stuff yourself, unless you are a massive business that can afford to, or it's the core of your business.
lhnz··on Don't build AI products the way everyone else is doing it
I think this is exactly the kind of issue that the startup Klu [0] is trying to solve. Not everybody should be building their own custom toolchains for finetuning/prompting AI models and there is really quite a lot of work involved in data management/evaluation.

There's definitely a lot of value in adding some AI features into your applications, but if it's not your core business you shouldn't be spending a lot of your time building a toolchain to do so.

[0] https://klu.ai

lhnz··on Ask HN: Who wants to be hired? (November 2023)
SEEKING WORK | LONDON, UK / REMOTE WORLDWIDE

Take a read of my blog or GitHub if you wish to quickly gauge my skillset: https://sebinsua.com

I'm a full-stack product engineer. I usually program in TypeScript, Python and Rust but have hundreds of hours of experience using a wide range of languages and technologies. I am an expert within the Node.js and React ecosystems, where I have 12 and 9 years of experience respectively.

  Location: London, UK

  Remote: Yes

  Willing to relocate: No

  Technologies: TypeScript, JavaScript, Python, Node.js, Next.js, Rust,
                SQL, PostgreSQL, React, HTML, CSS, Docker, Git,
                Kubernetes, Linux, Jupyter, FastAPI, Pandas, NumPy,
                Machine Learning, Large Language Models (LLMs), etc.

  Résumé/CV: https://sebinsua.com/resume.pdf

  Email: me@sebinsua.com

  GitHub: https://github.com/sebinsua

  Gists: https://gist.github.com/sebinsua

  LinkedIn: https://linkedin.com/in/sebinsua/
lhnz··on Ask HN: Freelancer? Seeking freelancer? (November 2023)
SEEKING WORK | LONDON, UK / REMOTE WORLDWIDE

Take a read of my blog or GitHub if you wish to quickly gauge my skillset: https://sebinsua.com

I'm a full-stack product engineer. I usually program in TypeScript, Python and Rust but have hundreds of hours of experience using a wide range of languages and technologies. I am an expert within the Node.js and React ecosystems, where I have 12 and 9 years of experience respectively.

  Location: London, UK

  Remote: Yes

  Willing to relocate: No

  Technologies: TypeScript, JavaScript, Python, Node.js, Next.js, Rust,
                SQL, PostgreSQL, React, HTML, CSS, Docker, Git,
                Kubernetes, Linux, Jupyter, FastAPI, Pandas, NumPy,
                Machine Learning, Large Language Models (LLMs), etc.

  Résumé/CV: https://sebinsua.com/resume.pdf

  Email: me@sebinsua.com

  GitHub: https://github.com/sebinsua

  Gists: https://gist.github.com/sebinsua

  LinkedIn: https://linkedin.com/in/sebinsua/
lhnz··on The Twelve-Factor App (2011)
It's pretty bad that, more than a decade after this was released, there are detractors for some of the guidelines but yet nobody is able to point to a singular more modern set of principles. A lot of our advice is now nuanced and spread out all over the internet...
lhnz··on Ask HN: Who is hiring? (October 2023)
It's good for some people but really it just means that if you have already put down roots in a location like San Francisco, New York or London then you'll have to look elsewhere for a job that pays more.
lhnz··on X rolls out new ad format that can't be reported, blocked
I didn't mean that literally zero users have left Twitter. But, I do not recognise its replacement by other services (Threads, Bluesky or Mastodon).

Obviously there are people that have cut social media out of their lives, and there are people that belong to Mastodon servers and so on, but this is a small subset of users.

lhnz··on X rolls out new ad format that can't be reported, blocked
Despite continuously making very unpopular moves, I can't see how Twitter loses its users. In fact, I think it will exist long after usage on Facebook, Instagram, Snapchat and TikTok falters -- because, it is less replaceable and less age-demographic dependent.

I also think that advertisers will return when budgets increase again. There aren't that many other good options.

In the long term, I think it goes nowhere and returns to profitability and a higher valuation.

lhnz··on Google User Data Is Police's Top Shortcut for Solving Crimes

  > Location data can easily be fabricated.
True. We are getting into territory that might also consider: what type of mobile device do you use, has it been jailbroken, and are there signs of GPS spoofing in the location history?

But, in general, if you have a digital trail which corroborates with other digital or real-life trails, you have something which does have weight as evidence.

lhnz··on Google User Data Is Police's Top Shortcut for Solving Crimes
I'm out-of-steam with this argument but making a probability-derived argument based on a cherry-picked agreement rate of 90% instead of an empirical argument that considers the legal deliberation process and culture that produced this agreement rate isn't persuasive to me.

Kalven-Zeisel specifically argues "Disagreement rates were no higher when the judge characterized the evidence as difficult than when the judge characterized it as easy, suggesting that the disagreements were not produced by the jury’s inability to understand the evidence." It seems impossible to mathematically tease out how much of the agreement comes from a shared understanding of the reasonable doubt standard vs other facts from the case -- either way, the idea that juries generally ignore reasonable doubt and err towards convicting people is not evidenced.

lhnz··on Vivaldi on iOS
Doesn't Kagi for iOS support web extensions already?
lhnz··on A Database for Game Engines
I recently saw this (https://twitter.com/AndrewChenE/status/1709000257747739046) which made me wonder whether we're close to being able to build 3D games in the browser using well-known web technologies like React.

However, in practice, if you want to build a cross-platform game, what is a reasonable game engine that you can use? I have an idea for a top-down 3D game and since I've never done this before don't want to waste too much of my time getting started. Is there anything that uses Rust, or is that a bad idea and I should look for something that uses a scripting language and something more well-known like C++?

lhnz··on Google User Data Is Police's Top Shortcut for Solving Crimes

  > I see no advantage to sharing the data in the cloud, vs just storing
  > it locally in your phone.
You have to look at this from the perspective of the reliability and integrity of digital evidence.

If the data is stored locally it could be argued that it's possible that you have tampered with it, particularly since you possess the skill set necessary to do so. On the other hand, if it is stored on the cloud, it's likely that you won't have the level of access required to tamper with the data, and that there will be an auditable transaction history with information about the identity of anybody that made changes. Google will be able to provide information to the court on the level of security that applies to the data -- in fact, data from Google services is already known to be legally admissable assuming it passes relevance, reliability and integrity tests.

Of course, you might be able to create something similar locally, but it's more complicated and untested: you'd need to be able to prove that you don't have the ability to retroactively amend the data (write access, encryption keys, etc).

I agree that it would be nicer if: the digital evidence wasn't accessible by others and was also reliable and with integrity protection. Maybe an option would be to encrypt it with your public key and write it into a public blockchain?

← PreviousPage 2 of 34Next →