HN user

d--b

9,528 karma

Software developer + wannabe screenwriter.

Posts77
Comments2,862
View on HN
news.ycombinator.com 1mo ago

Ask HN: How do you avoid / get out of LLMs local minima?

d--b
3pts3
www.theguardian.com 2mo ago

Musk vs. Altman lawsuit over OpenAI starts today

d--b
8pts0
news.ycombinator.com 4mo ago

Ask HN: Quick technical questions about LLMs

d--b
2pts0
news.ycombinator.com 8mo ago

Ask HN: Do you know of a AI company that's using user's clients to request stuff

d--b
1pts0
en.wikipedia.org 9mo ago

The Right to Be Lazy (1883)

d--b
3pts0
www.theguardian.com 9mo ago

A bit of nostalgia: Space Harrier game is 40

d--b
7pts1
news.ycombinator.com 1y ago

Ask HN: Are Trump's tariffs affecting freelancing from abroad?

d--b
2pts1
news.ycombinator.com 1y ago

Ask HN: A friend has brain cancer: any bio hacks that worked?

d--b
157pts364
www.get-aerial.com 1y ago

What's Up Doc? Introducing doc-as-dashboard

d--b
1pts2
www.get-aerial.com 1y ago

Aerial – hand drawn live dashboards

d--b
2pts0
news.ycombinator.com 1y ago

Ask HN: Product to improve monitoring and documentation: a good idea?

d--b
4pts5
news.ycombinator.com 1y ago

Ask HN: Most useful 3D printed items?

d--b
3pts3
www.theguardian.com 2y ago

The delicate art of treating ADHD

d--b
4pts0
twitter.com 2y ago

Musk to OpenAI: Change your name to ClosedAI and I will drop the lawsuit

d--b
4pts1
techxplore.com 2y ago

OpenAI turning to designer Jony Ive for 'AI iPhone'

d--b
2pts0
news.ycombinator.com 3y ago

Ask HN: I Think I Need a New Calendar App

d--b
26pts37
news.ycombinator.com 3y ago

Ask HN: Why can't I reply to comments after a while?

d--b
7pts7
news.ycombinator.com 3y ago

Tell HN: Half of current “Show HN” posts are about AI

d--b
3pts1
en.wikipedia.org 3y ago

Chinese Room

d--b
4pts1
news.ycombinator.com 3y ago

Ask HN: Did GPT develop some form of consciousness?

d--b
1pts2
observablehq.com 3y ago

The benefits of JavaScript for data analysis

d--b
1pts0
news.ycombinator.com 4y ago

Ask HN: Is the NFT Craze Over?

d--b
5pts6
news.ycombinator.com 4y ago

Ask HN: Do you upvote bad posts so that more people read your criticism?

d--b
4pts4
github.com 4y ago

Show HN: Procrastinope – insert your todo list into your procrastination site

d--b
4pts2
www.jigdev.com 4y ago

Show HN: Jig – Live coding your business dashboards (now with demo video)

d--b
31pts8
edeboursetty.substack.com 4y ago

End User Data Analysis: The Good, the Bad, and the Future

d--b
2pts0
edeboursetty.substack.com 4y ago

A History of BI in 6 crappy drawings

d--b
2pts0
edeboursetty.substack.com 4y ago

Interfaces are contracts between people

d--b
1pts0
news.ycombinator.com 4y ago

Ask HN: Is ProductHunt Worth It?

d--b
99pts66
www.jigdev.com 4y ago

Show HN: Jig – a tool to define, compute and monitor metrics

d--b
74pts24

England has Shakespeare, Spain has Cervantes, Italy has Dante, and Russia has Pushkin

Author is conveniently avoiding mentioning France, which does not have a major figure weighing on the others that much. Hugo is probably the most homeric of the classic French litterature, but he’s much more recent and nowhere nearly as dominant as Shakespeare in England or Cervantes in Spain (which borders on the ridiculous in both countries imo).

It’s a better situation not to have an Homer but a constellation of good stuff.

You’re saying that if the LLMs were right, humans would have been correct in trusting the machine for things they didn’t know?

The point is that LLMs aren’t right, and the people who took the test were probably reminded of that.

Would people have trusted the textbook you’re mentioning if there was a big red warning on each page that said “this book may contain errors”.

The willingness to trust AI even though it may be wrong and even though there’s money on the line is intersting enough as a study imo

While it is easier to think of large numbers in terms of logarithms, cause it makes unfathomly large numbers much more palatable, I disagree that it makes scales more intuitive.

1 billion is a very large number, and thinking of it as 10^9 make it seem smaller.

1 trillion is "just" 3 orders of magnitude above 1 billion, and "only" 9 orders of magnitude less than the number of atoms in a mole.

I don't know the answer to making the mind understand scale. I don't think things like "it 's about 2000 football pitches" help either. I don't think "a billion is the number of cubic milimeter in a cubic meter" work either. I don't think the logarithm based "zoom visualization" work either. I just think the brain just cannot picture what those numbers mean. We're not wired to understand those things very well, just as we aren't wired to work with 4+ dimensions

I am somewhat surprised not to find any Jodorowsky in there.

Maybe the incal world doesnt really qualify as punk and more like regular scifi or perhaps bordeline new-agey, but it doesn’t strike me as fundamentally different from Akira in story style.

I think OP is mainly right about this, but the tone of the article is going to annoy people very much. I believe many of us have done both the self-serving homemade framework and the industry standard framework. And many of us know the pros and cons of both approaches.

I don’t like react very much and dislike the other ones even more. Yet, it’d be stupid to say you can’t make them work. They come with a large community and pckages and solve a ton of problems that a homebrewed framework doesn’t even conceive of. And you can just start writing the stuff you need right away.

On the other hand, a homemade framework doesn’t require the extra overhead of knowing the framework, and maintaining it overtime. You don’t need to track down why the code gets run a gazillion times more than it needs to, you can debug through your events, etc.

They’re just different ways of doing things.

But mainly: react is a pain in the butt but if time is your constrain, it should get you to where you want to be faster than if you’re not using it.

Resetting Xbox 16 days ago

Today, in some parts of the company, work passes through as many as 14 layers of management.

We will reduce management layers to no more than 5, and where possible, 3

I don’t understand why they need to make a big announcement when what they’re doing is cutting the stupid management structure. Everybody hates cuts, except those that cut middle management.

They should have titled this “xbox cuts middle management: Less burocracy, faster delivery, more original games, more money for everyone involved, yay!”

There is a process in place that’s meant to capture a certain number of potential problems. I didn’t make that process. The people who are making drugs safe designed the process. There is never zero risk of a treatment behaving badly, of course but when a drug gets fast tracked and doesn’t go through the regular approval process, it just hasn’t been proven to be safe by the regular standard of what experts deem safe.

It’s not very complicated.

trials ok => drug most likely ok

trials not done => we don’t really know.

By late 2020, when they got approved, the vaccines were not scientifically proven safe for mainstream use. No other mRNA vaccine had been through all the trial stages, and certainly not those COVID ones.

Could the vaccines have side effects that became visible after 6 months? Yes and we couldn’t have known that they didn’t.

Could the vaccines have side effects on people with rare conditions? Sure, and we couldn’t have known that either.

My point is that in 2020, the decision to approve the vaccines and pretty much force everyone to get it was a risk tradeoff. It was way more risky to let the disease continue spreading and mutate than it was to release the vaccines. mrna vaccines had been in trials and there was no reason to believe they could have been harmful. But the reality is that we just didn’t know. Biology is complex enough that you can’t just assume everything will be fine without proper testing. And what we deem proper testing is a process that these drugs hadn’t gone through.

I happily got vaxed in early 2021, and did it again 4 times , so I was willing to trust the tradeoff.

But ignoring that it was a tradeoff and hiding behind a sign that says “science” is just taking people for dummies.

Online privacy is already an illusion.

I mean, your ISP knows your IP, the government can know your IP, Google knows your IP, Meta knows your IP, all the websites you log on know who you are and what you're IP is, and forward your data to 3rd parties, many of these connect the dots between the various websites you visit.

If there is one benefit to true ID verification on the internet, is that people won't feel as if they're browsing things anonymously while they're not.

The price of a meme coin reflects only the collective belief of meme coin holders that they will be able to sell to a greater fool

Author is at times a little too emphatic, but he has some sentences like this one that are really efficient in conveying the idea.

yes it's a bit of a mystery to me as well, especially knowing that most of them are scams.

There are 2 things I don't understand.

1. Most online ads are illegal in EU, cause it's illegal to lie/mislead in an ad in the EU. The Tai Chi ads, the water pressure thing or the breezamax scams are all I see on Youtube, yet, they're obviously fake, and still running strong. How is it that they're not taken down?

2. The tai chi ad campaign must be costing millions of dollars. Everyone I know is seeing these fifteen times a day. Is anyone actually paying for that crap? Shouldn't Youtube increase the price of ads, so that they make more money and actually push scammers off the platform?

It doesn't make any sense

Young children can acquire absolute (perfect) pitch — but adults cannot. The window closes around age 6.

What?

That doesn’t sound true at all.

Jude Kofie started the piano at the age of 8. But music prodigy apart, I defintely don’t have perfect pitch right now, but I am very confident that I could with some training.

Like, when I sing a song without music, I usually am very close to the right tone. Mostly because I intuitively know how my voice is supposed to feel (the vocal chords move differently based on the tone).

And I can clearly ear the different “colors” between different scales.

And when I play the guitar a lot, after a month or so I start to be able to know where a note I hear is on the guitar.

A CORS protected endpoint tells YOUR BROWSER not to let YOU access its content if the website you’re browsing from is not whitelisted.

It’s confusing because unlike most security features, it’s meant to protect the users from themselves. The risk comes from a combination of users being allowed to visit malevolent sites and browsers letting all websites do a lot of random stuff, including making 3rd party requests with cookies and private stuff

GEMINI 3.1 LITE SAYS A FRENCH DIPLOMAT WHO SERVED AS THE AMBASSADOR OF FRANCE TO THE REPUBLIC OF THE CONGO (BRAZZAVILLE) FROM 2017 TO 2020.

er. okay. The good thing about this test is that I am the only person in the world with my full name, and I know all the people with my last name (about 30-ish people). None of us are ambassadors, none of us are related to Congo in any way.

2 things:

1. Like everything, apts degrade over time, maintenance usually happens in spikes, so the state of the place goes down until someone pays for the new paint job

2. It depends on the market, in highly attractive places, owners don't care about keeping tenants, so they'll let the place fall apart, until they can't get anyone anymore. Then they'll put in the price to put the place back in order and start again.

It’s the GPUs, not the weights that are the key.

As long as these models require a lot of computing power, the best models open source or not will be served by corporations who can afford the infra.