HNHacker News
TopNewBestAskShowJobs

FridgeSeal

7,095 karma · joined March 20, 2017

Opinions are my own etc etc etc
submissionscomments
FridgeSeal··on Dogs may have domesticated themselves because they liked snacks, model suggests
My parents dog OD only interested in food if:

- the treats you have are good-enough (better have delicious chicken, otherwise, no dice)

- she hasn’t decided that there isn’t something else she’d rather be doing

- hours much of a process she feels like being today.

You can motivate her with food only on her terms. Conversely, my partners dog would commit crimes just to be near the cheese packet.

FridgeSeal··on Show HN: I made a site to tell the time in corporate
No stop it, you’re stressing me out just reading it.
FridgeSeal··on MongoDB acquires Voyage AI
These days almost all the alternatives do 90% of what they cover in a way that’s either using your pre-existing tech stack, or just net-better.

There’s probably some extremely specific niche use-cases where it works well, but I suspect they’re pretty few-and-far-between.

FridgeSeal··on MongoDB acquires Voyage AI
Because some devs and teams prioritise “get to prod” above literally all else.

Maintainability? Secondary. Security? Secondary. Data-integrity/correctness? Secondary.

FridgeSeal··on Vine: A programming language based on Interaction Nets
This is really, really cool!

I don’t pretend to understand all the theory for nets stuff, but the premise seems super cool and I love languages that have experimental and unusual ideas!

Also liked the dedicated section about the compiler design: the description of the different passes is cool.

FridgeSeal··on When your last name is Null, nothing works
> Even things like the automated customs gates work fine.

I would like to think (Perhaps naively?) that these systems key off your passport number/ID (which is by construction, not subject to these problems) to deliberately side-step issues like this.

FridgeSeal··on Nuclear fusion: WEST beats the world record for plasma duration
“Well, created one nice one, and one desolate, irradiated, lifeless shit-hole! My work here is done!” Haha
FridgeSeal··on You're not a senior engineer until you've worked on a legacy project (2023)
Nothing quite like suffering through your own decisions to make the lesson stick!
FridgeSeal··on Did Semgrep Just Get a Lot More Interesting?
To be fair, that also could be because the pandas API is…not great. When I used to write lots of python and pandas I basically always had the pandas docs open intuitive and memorable it is not.
FridgeSeal··on Did Semgrep Just Get a Lot More Interesting?
We all keep making this joke, and one day someone’s going to try it. I have a feeling it’ll be moderately successful too.
FridgeSeal··on We are destroying software
I could finally get all the skiing in that I wanted to with time off like that!
FridgeSeal··on We are destroying software
> Don’t reinvent the wheel!” But reinventing the wheel is how you learn how things work, and is the first step to make new, different wheels. >We are destroying software pushing for rewrites of things that work.

> We are destroying software by jumping on every new language, paradigm, and framework.

Whilst I largely agree with the premise of the post, some of these points feel a little bit dissonant and contradictory. We can have stability and engineering quality _and_ innovation. What I think the author is trying to target is unnecessary churn and replacements that are net-worse. Some churn is inevitable, as it’s part of learning and part of R&D.

FridgeSeal··on We are destroying software
> Consumers want more features

I’m not even sure if this is true anymore. We got new features foisted on us most of the time.

FridgeSeal··on We are destroying software
As a group, we have trained many, many programmers out of even considering performance with the proliferation of quotes like “premature optimisation is the root of all evil” ands ideas like “who cares, just get a faster computer/wait for hardware”.

Premature optimisation is bad, but there’s now so many devs who don’t do _any _ at all. They don’t improve any existing code, they’re not writing software that is amenable to later optimisation, inefficient architectures and unnecessary busywork abounds.

Are we surprised that years of “product first, bug fixes later, performance almost never” has left us with an ecosystem that is a disaster?

FridgeSeal··on Asahi Linux lead developer Hector Martin resigns from Linux kernel
I saw the social media post, and it didn’t even register as “trying to start a brigade” IMO.

I read it as “look at this shit” and general frustration with the likes of DMA-guys behaviour.

FridgeSeal··on Asahi Linux lead developer Hector Martin resigns from Linux kernel
How is that literally any different from C code?

There’s 2 kinds of scenarios here:

- Rust code involved: C code change breaks Rust. ~Send~ ~an~ ~email~ ~and~ move on.

- C code involved: inform other C devs, either work with them or wait for them to fix.

The way I see it, either way, the “Rust stuff adds unacceptable amounts of workload” doesn’t really check out.

Edit: turns out he doesn’t even need to inform the Rust devs.

FridgeSeal··on Hytradboi 2025 Program
Really keen for this! The talks from the last one were really interesting.

The no-corporate-sponsors fees like a good model too, the previous talks were all detailed, technical talks aimed at the same audience, nice to not be sold to.

FridgeSeal··on Analyzing the codebase of Caffeine, a high performance caching library
Google-Adjacent OSS repos are the worst for this!

Passive-Aggressive Stalebot marches into the middle of an _active_ thread and announces that the issue is now stale/dead and everyone should go away.

I get that on large projects there’s a non-trivial percentage of issues which amount to “I’m holding it wrong, didn’t actually read the log messages, or the manual, fix it for me pls” which are just unhelpful noise. However more often than not they take every other thread- including important ones, with them.

FridgeSeal··on Hydro: Distributed Programming Framework for Rust
This looks to be a degree more sophisticated than that.

Authors in the comments here mention that the flo compiler (?) will accept-and-rewrite Rust code to make it more amenable to distribution. It also appears to be building and optimising the data-flow rather than just distributing the work. There’s also comparisons to timely, which I believe does some kind of incremental compute.

FridgeSeal··on OpenAI O3-Mini
I think it’s success in spite of branding, not because of it.

This naming scheme is a dumpster fire. Every other comment is trying to untangle what the actual hierarchy of model performance is.

FridgeSeal··on OpenAI O3-Mini
OpenAI chatGPT Pro Max XS Core, not to be confused with ChatGPT Max S Pro Net Core X, or ChatGPT Pro Max XS Professional CoPilot Edition.
FridgeSeal··on OpenAI says it has evidence DeepSeek used its model to train competitor
I find it interesting that so much discussion about “LLM’s can do some of our work” is centred around “are they intelligence” and not what I see as the precursor question of “are we doing a lot of bullshit work?”

My partner is in law, along with several friends and the amount of completely _useless_ work and ceremony they’re forced to do is insane. It’s a literal waste of their talent and time. We could probably net most of the claimed AI gains by taking a serious look at pointless workloads and come out ahead due to not needing the energy and capital expenditure.

FridgeSeal··on OpenAI says it has evidence DeepSeek used its model to train competitor
No you don’t understand, AI is “dangerous” and only him and his uber rich billionaire mates should get to control it!
FridgeSeal··on OpenAI says it has evidence DeepSeek used its model to train competitor
No no no you don’t understand, the models will magically overcome issues and somehow become 100x and do real AGI! Any day now! It’ll work because LLM’s are basically magic!

Also, can I have some money to build more data centres pls?

FridgeSeal··on Composable SQL
I think you’re throwing the baby out with the bath water here a bit.

I agree that stored-price are rarely, if ever the solution.

I do however think that having something that enforces certain constraints across all your data, and forces you to handle schema evolutions reasonably over time is really powerful. Database implementations are (mostly) full of hard-fought correctness decisions, and I don’t think most devs are in a position to “whip out” an implementation that cares for a number of these on-demand.

We had the whole NoSQL thing, we saw what when we threw most of this stuff away.

Is sql good enough? For many problems yeah. Is it _really_ good enough? Nah probably not, we could do a lot better, but I think a better design looks more like a better sql-like language and less like a full programming language.

FridgeSeal··on A layoff fundamentally changed how I perceive work
We also start, or encourage starting work quite early on in our lives, and so it naturally grabs a place in people’s existence in their formative years as “a thing that they do”. Is it any surprise then, that it naturally ends up becoming at least a non-trivial part of people’s sense-of-self?
FridgeSeal··on DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL
Whenever I use it, it just seems to spin itself in circles for ages, spit out a half-assed summary and give up.

Is it like the OpenAI models in that in needs to be prompted in extremely-specific ways to get it to not be garbage?

FridgeSeal··on DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL
I didn’t really have this experience.

Every time I tried it, the thinking mode would spin for years, it’d send itself in a loop, not do anything I instructed in the prompt, and then just give a weird summary at the end.

Claude models correctly parsed the prompt and asked the follow-up questions.

Edit: tried it a few more times. Without the “R1” mode enabled it genuinely just restated the problem back to me, so that’s not ideal. Enabling R1 and pointing that out has sent it into a loop again, and then produced a wildly-overcomplicated solution.

FridgeSeal··on Explainer: What's r1 and everything else?
It’s basically the line for all the AI-hype people: “all the problems are going away!”, “soon it’ll all magically make things exponentially good-er-er!”
FridgeSeal··on Larry Ellison: vast AI surveillance can ensure citizens are on best behavior (2024)
I wish they would go and live in their own private world and leave the rest of us alone, instead of trying to inflict something that the Stasi-would-be-proud-of on the rest of us.
← PreviousPage 12 of 34Next →