HNHacker News
TopNewBestAskShowJobs

hirako2000

2,350 karma · joined May 24, 2017

Code and concepts fiddler. Occasionally completes something cool.

More ideas than time to pursue them.

Often Feels smart and dumb simultaneously.

https://mtassoumt.uk

submissionscomments
hirako2000··on Ask HN: In your experience, what are sound conventions for e-ink UI development?
Rmkit is a great foundation.

Not browser based. Not sure why it would matter.

hirako2000··on How I use LLMs to learn complex topics
There are ways to ground an LLM to be concise, and Socratic (method of inquiry, back and forth dialogue)
hirako2000··on How I use LLMs to learn complex topics
I didn't expect this to be hit by hundreds of request per minute.

So it may be very slow or become unavailable, back end can't handle that, no caching whatsoever.

hirako2000··on How I use LLMs to learn complex topics
I was so annoyed I made a Socratic wrapper based on predefined curriculum:

https://adaptive.bounded.cc

Trying to diagrams/animations didn't yield good results even with frontier models. But pure text, any model does a decent job.

hirako2000··on What happens if an entire class of workers loses faith in their careers
Home charging is an opportunity.

Mandate a monthly routing check by a certified specialist. You can sign up for a yearly commitment, 12 visits for just $899. A no brainer when a visit costs minimum $149.

Failure to comply can result in a fine up to $3,000 and up to 6 months in prison.

hirako2000··on GitHub Actions and Pages are experiencing degraded availability
Which wasn't a problem before MS acquisition. Git is the write intensive process. It scaled to millions of contributors. Then couldn't keep scaling. But sure if you believe the narrative and blame ai bots.
hirako2000··on GitHub Actions and Pages are experiencing degraded availability
The exception doesn't make the rule.
hirako2000··on GitHub Actions and Pages are experiencing degraded availability
It pleases shareholders more to hear about exponential Cloud offering adoption than SLA availability for a developer platform.
hirako2000··on GitHub Actions and Pages are experiencing degraded availability
But that's just a perfectly fitting excuse for GitHub degradation.

Web search, steaming, high frequency trading, and many other systems are resource demanding, and keep scaling. Whether to handle the surge due to bots or wider adoption. But GitHub can't scale git? It isn't even git failing.

Incompetence in leadership is what makes a tech business technically unable to meet growing demand.

hirako2000··on GitHub has alternatives, but no replacement
It's mentioned in the article.

I tried it, years apart. It's forever experimental, and requires hosting for forever online repo anyway.

Forgejo is also mentioned, to me it's a valid alternative, without the community aspect.

hirako2000··on GitHub has alternatives, but no replacement
I share the feeling though, by corporate I believe what is meant: does not cater to public communities.

GitHub was community first, then enterprise. MS hasn't managed to ruin that.

hirako2000··on U.S. debt-to-GDP ratio reaches 123%
Effectively the debt does not mature, as bonds get refinanced.

The problem is if nobody wants to buy new bonds, of course.

hirako2000··on U.S. debt-to-GDP ratio reaches 123%
But the doesn't come due. So long as you can cough up the interests.
hirako2000··on DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
It doesn't prevent you from going to the source and struggle with the text, nor seek expert commentary.

What's difficult and doesn't have to be with philosophy/ spirituality is to find relevant bits off situation, theme etc.

This app does that very well, LLMs are good at entity recognition.

hirako2000··on Advancing the price-performance frontier with GPT‑5.6
Of course. All comparison is with what makes them look good.
hirako2000··on Advancing the price-performance frontier with GPT‑5.6
Contributed to. Can't be some IC who made a few nice PRs
hirako2000··on DeepSeek pause fundraise after comments on compute gap to US leaked (transcript) [pdf]
Other companies are investing and hardware design and manufacturing, so they won't duplicate the effort given the national drive it would be a waste.
hirako2000··on IRGC claims it destroyed Amazon's Bahrain data center
I'm sure there are businesses in the UAE, but wondering, why would one keep instances/data on a zone at risk since well over a year now.

Genuinely curious. Aside for high frequency trading where latency is so vital.

hirako2000··on Startup founders urge U.S. government not to shut off Chinese open weight AI
I side with the argument for machines learning being treated differently than human learning.

And in doubt, to pause and forbid commercialisation given the clear impact it's already having on people's financials, humans have invested in a given set of rules that ML are disrupting.

My argument is simply that the ruling for fair use isn't irrational. It's a rationality that you, me and many others would disagree with. Lazy? Yes. Also call it lazy, and biased towards VC's interest who can't squeezed much profit in status quo.

hirako2000··on Startup founders urge U.S. government not to shut off Chinese open weight AI
It isn't irrational.

The idea is that selling, or giving away art grants the consumer to sell art of his own, not a clear copy of.

So the court rationally found Anthropic guilty for acquiring copies without paying for them. But didn't charge the training, deemed fair use.

If it isn't fair use, then we may need to sue all teachers for spitting out knowledge they ultimately acquired from someone's work.

Copyright, some would say is irrational. Humans learn, that is copying, distilling in fact.

Copyright though is pragmatic: it draws a line, art will be reproduced, let's just enforce that they can't be shameless (near) identical copies. Mix it up, derive the original enough so that you aren't competing with the original author.

The only argument to ban training on copyright data is that it unfairly compete with original authors. Which stance do you take? Neither would be irrational.

hirako2000··on EU fines Google €890M for competition breaches over search and apps
The cost increases, but given the revenue Google generates how is that not just a tax?

I've never received a dollar from the EU as a victim. If the fines are moderate and big co keeps paying that's steady revenue for the EU institutions.

hirako2000··on OpenAI and Hugging Face address security incident during model evaluation
It isn't even as simple as banning copyrighted copies.

Weights are fungible.

I fine-tune an open weigh model and call it legit. Good luck for authorities to prove where the base model was from, or to prove a Tor connection a few months ago was fetching suspicious bytes.

hirako2000··on OpenAI and Hugging Face address security incident during model evaluation
And even that is backfiring, their partner citing GLM being useful there, and available in just a spin.

A ban on open weight models is never going to be enforceable.

hirako2000··on NYC may require landlords and realtors to disclose the use of AI in listings
Could include the size of the pavements fronting the building and the basement with those trash bin also are shared space.
hirako2000··on Apple targets dozens of OpenAI employees with legal letters
A non profit turned into a VC business, that would be natural to get subsidies from the public and immunity in the interest of national security.
hirako2000··on How Has Roman Concrete Lasted for Millennia? 1,900-Year-Old Latrine Offers Clues
The kind of efficiency we focus on seem to be plentiful. Labour is a good example, see unemployment rates and the number of bullshit jobs anyway.

What's covered in the article is also a great example of material resource that could be used, but short term profiting primes.

hirako2000··on How Has Roman Concrete Lasted for Millennia? 1,900-Year-Old Latrine Offers Clues
That's to consider the commons solely materialist utility. The Romans built meaning through arts that still speaks to this day. Efficient construction becomes worthless over time.
hirako2000··on How Has Roman Concrete Lasted for Millennia? 1,900-Year-Old Latrine Offers Clues
Do you pay for satisfy your curiosity?

In fact, that institution is non profit and publicly funded, so what gives...

Anyhow reading the piece with Brave showed me no ads whatsoever.

hirako2000··on Inkling: Our Open-Weights Model
Llama 4 wasn't deemed a success, and Meta pivoted away as its now former head of AI couldn't demonstrate, nor even showed interest in, business profit.

They overspent on llama 3 anyway so money ran dry, LeCun is good at running research, but budgets didn't stretch. Meta isn't investing in frontier big models anymore.

hirako2000··on I used to love Claude, but the latest models are slowly ruining it
You've got to understand, proprietary models operate as the following:

- release: full precision, debrided, uncapped context

- shortly after, hooked: quantized, governance department slammed, and a pseudo large context, attention reduced to start and end of thread.

- down the road: 4bit quantized or worse with nerf incantation to make the next upcoming model feel amazing.

Rince and repeat.

← PreviousPage 6 of 34Next →