HNHacker News
TopNewBestAskShowJobs

ex-aws-dude

1,128 karma · joined December 17, 2025

submissionscomments
ex-aws-dude··on Tesla to wind down Cybertruck production
I'm guessing its a sport bike though, dirt bikes are way taller
ex-aws-dude··on Tesla to wind down Cybertruck production
"You never see a whole family in a cybertruck, its always just one man"
ex-aws-dude··on Tesla to wind down Cybertruck production
It totally depends on your use case

One of the reasons I love a pickup is that it separates the cargo from the interior

A lot of times I have oil/gas/chemicals/something dirty in the bed which I would not want in the cabin space

Also it depends do you care about height of cargo or not?

For example a minivan can fit plywood but not a dirt bike

ex-aws-dude··on Tencent Hy4
> By Jevons' paradox

Jevons Paradox is not a rule its just a sometimes observed phenomenon

It says nothing one way or the other whether that would happen or not

ex-aws-dude··on MCP Explained
That's why its better for the person who just learned something to write the docs (checked by the expert though obvs)
ex-aws-dude··on sshole
No the biggest lesson of all this AI stuff to me is that we are just not smart enough as programmers

If you think about it that’s the whole idea of neural nets

We aren’t smart enough to design the function we want by hand as programmers so we need to learn it through weights

This lesson just keeps recurring in many different ways

ex-aws-dude··on Ask HN: What's your current AI agentic setup?
This thread is super biased by the power users, they are the most excited to share
ex-aws-dude··on Progress on provable inference for large language models
> Some proposals to regulate and pace AI development rely on monitoring hardware and how it is used. For example, a provider might be required to devote its hardware to inference rather than to training a powerful new model. The provider might try to cheat by serving cheaper models or using compute in an unexpected way, freeing up resources for training. How can a regulator reliably ensure that the hardware is used for inference?

Is this even a real problem? Forcing a company to only do inference?

Sounds like Anthropic's dream scenario

ex-aws-dude··on OpenAI seems to be silently re-enabling model training on Codex chats
Why so aggressive? Other people have shared the same thing https://news.ycombinator.com/item?id=49643556
ex-aws-dude··on Handles are the better pointers (2018)
Also something not discussed is that you kind of want handles anyway for serializing objects that reference eachother
ex-aws-dude··on Postman silently auto-installs AI plugins and attempts to share all API data
I'm not really a vibe-coder but it seems like the easiest possible tool to vibe-code

Or just give your agent JSON and tell it to use curl

ex-aws-dude··on Postman silently auto-installs AI plugins and attempts to share all API data
As someone who hasn't used it since like 8 years ago that reads like satire
ex-aws-dude··on Handles are the better pointers (2018)
I’ve worked on game engines and what he’s describing is still very common

I don’t know about outside that world though

ex-aws-dude··on The 'Astra can read and write weights to improve Qwen' post is misleading
My main takeaway is that HN had no idea the field of interpretability existed until now

I don't know why because Anthropic has had a lot of blog posts about internal work on it

ex-aws-dude··on The 'Astra can read and write weights to improve Qwen' post is misleading
Are you kidding? If anything pro-AI hysteria is way stronger here, just look at how many upvotes the original got

Even any small model update a company puts out is guaranteed top of HN

ex-aws-dude··on The 'Astra can read and write weights to improve Qwen' post is misleading
They might even discover some way to program the weights proportional to an error signal

Now that would be crazy

ex-aws-dude··on Astra can read and write weights to improve Qwen
I don't think its that crazy that you could look at activations and tweak weights directly to get a certain outcome, but isn't the question whether you can do that without side effects or degrading/breaking other parts of the model?

If I want to do that in a generalizable way that preserves performance across all domains then am I not just back to training again?

ex-aws-dude··on Astra can read and write weights to improve Qwen
I've been finding that too with these models

For example pointing them at a crash dump they manage to piece together and extract information I would not even try to extract by just looking at the raw memory

ex-aws-dude··on Which hobby do you think is most diverting and pleasurable?
Anything where you are out in a forest
ex-aws-dude··on Show HN: HackerBooks – HackerTube, but for books mentioned on HN
Pfft that is simply a special case of HackerRecommendationSiteGenerator.com
ex-aws-dude··on Mathematics has changed. Let's act like it
Ive noticed that too, I don’t know why people talk like we need to figure this out right now

Like we can even know what the future of math is this early

It’s not like you can just declare that up front, it’s something that happens naturally when people spend time using the tools

ex-aws-dude··on AI delivers 'zero financial payoff' to most companies, Apollo economist says
You don’t get it, what’s the actual profit gained to the company from that?

That’s the whole distinction

ex-aws-dude··on HackerTube: YouTube built from videos recommended by HN
You’re not wrong, sometimes even a tweet by these people will be front page
ex-aws-dude··on We're releasing the weights of Mistral Large 4
Duh? If someone uses it to try hacking you its still 96.6% refusal
ex-aws-dude··on Building an AI-native engineering org
Are they even making 3M per month?

Isn't there a point where you're shipping too fast that the product hasn't even caught on enough yet to know what to build?

I looked at some of the games and it just looks like a Roblox clone

ex-aws-dude··on Binance-hacker-1029 – $5.6B stolen from Binance
That is the tech bro circle of life

Invent Notion then keep adding features enterprise customers want until you're re-inventing JIRA/Confluence

Who left this stupid fence here?

ex-aws-dude··on We're releasing the weights of Mistral Large 4
It is kind of comical how things that previously would have been obviously true are now only statistically true

"Our product won't hack you 96.6% of the time"

ex-aws-dude··on Toxic experts (2017)
I consider going online in general as going into the pig pen at this point

I’ll still go in to see what’s happening but don’t forget where you are and don’t take it too seriously

ex-aws-dude··on The Grid, a factory-scale robot in Massachusetts that builds PCBs on demand
"The Grid. A digital frontier."
ex-aws-dude··on Ask HN: What do you believe or not believe about AI?
Not enough people are talking about the 1st point

Learning that it’s all just computation has some uncomfortable philosophical implications, perhaps religious too

Page 1 of 13Next →