HNHacker News
TopNewBestAskShowJobs

risyachka

1,091 karma · joined November 7, 2017

submissionscomments
risyachka··on Anthropic raises $13B Series F
>> The compute moat is getting absolutely insane.

how so? deepseek and others do models on par with previous generation for a tiny fraction of a cost. Where is the moat?

risyachka··on New research reveals longevity gains slowing, life expectancy of 100 unlikely
Natural one maybe.

Humans are biological machines. We know how to replace hearts with artificial ones that can last years. Soon they may start lasting decades.

We can replace many hormones with artificial ones.

I do not see any reason we can’t learn how to replace other organs and systems.

And in this case you may as well live 200+ years

risyachka··on I built a production app in a week by managing a swarm of 20 AI agents
So where is the app?
risyachka··on Scamlexity: When agentic AI browsers get scammed
Agenetic browsers is like a sealing tape fix of a high pressure water pipe - it should not exist.

If you want the agent to do things for you - there is literally zero reason to use a browser instead of an API.

Like 1 bulletproof API call vs clicking and scrolling and captcha and scam stores etc - how can this possibly be a good idea?

risyachka··on Making games in Go: 3 months without LLMs vs. 3 days with LLMs
This.

You may as well buy a shooter game starter pack or whatever that can save you >1year of coding, no llm needed.

Code is not a hard part.

Making mechanics fun and good assets is what is hard and takes forever.

Sure you can use llm to write a generic game, but its easier to find same game on github and just use that code, why would you write it again with llm.

risyachka··on The AI vibe shift is upon us
Why? Doesn’t sound different from any framework e.g. React.

Fixes one pain point good. Can’t really be applied to everything.

So just another tool, not a magic bullet like it is being marketed.

risyachka··on AI tooling must be disclosed for contributions
It should be. You didn’t write generated code, why should I spend my life reading it?

If you want me to put in the effort- you have to put it in first.

Especially considering in 99% of cases even the one who generated it didn’t fully read/understand it.

risyachka··on AI tooling must be disclosed for contributions
>> but maybe those don't need to be about "whether or not you used LLMs"

The only reason one may not want disclosure is if one can’t write anything by themselves, thus they will have to label all code as AI generated and everyone will see their real skill level.

risyachka··on How we exploited CodeRabbit: From simple PR to RCE and write access on 1M repos
Software industry really needs at least some guardrails/regulations at this point.

It is absurd that anyone can mess up anything and have absolutely 0 consequences.

risyachka··on AI-induced dehumanization (2024)
> Are we simply getting old and bitter?

Maybe, but it has nothing to do with change itself.

Change can be either positive or negative. Often it is objectively negative and can stay that way for decades.

risyachka··on Claude Sonnet 4 now supports 1M tokens of context
Pretty much my experience too.

I usually go to option 2 - just write it by myself as it is same time-wise but keeps skills sharp.

risyachka··on GPT-5
Considering how they hyped it up (eg. “Lol normies go about their day and have no idea whats coming etc”) they have to show some AGI level llm or stop overhyping their 2% improvements.
risyachka··on 6 weeks of Claude Code
I mostly referred to skills, not brain function itself.
risyachka··on 6 weeks of Claude Code
"mental strain" in a way of remembering/thinking hard is like muscle strain. You need it to be in shape otherwise it starts atrophying.
risyachka··on 6 weeks of Claude Code
Its all great until

>> I thought I would see a pretty drastic change in terms of Pull Requests, Commits and Line of Code merged in the last 6 weeks. I don’t think that holds water though

The chart basically shows same output with claude than before. Which kinda represents what I felt when using LLMs.

You "feel" more productive and you definitely feel "better" because you don't do the work now, you babysit the model and feel productive.

But at the end of the day the output is the same because all advantages of LLMs is nerfed by time you have to review all that, fix it, re-prompt it etc.

And because you offload the "hard" part - and don't flex that thinking muscle - your skills decline pretty fast.

Try using Claude or another LLM for a month and then try doing a tiny little app without it. Its not only the code part that will seem hard - but the general architecture/structuring too.

And in the end the whole code base slowly (but not that slowly) degrades and in longer term results net negative. At least with current LLMs.

risyachka··on 6 weeks of Claude Code
>> Review the code that comes out. It goes a long way.

Sure, but if I do 5 reviews for a task - in 99% of cases it is net negative as it is faster to DIY it at that point. Harder for sure, but faster.

risyachka··on 6 weeks of Claude Code
I bet it highly depends on the work you do.

It is very useful for simpler tasks like writing tests, converting code bases etc where the hard part is already done.

When it comes to actually doing something hard - it is not very useful at least in my experience.

And if you do something even a bit niche - it is mostly useless and its faster do dig into topic on your own that try to have Claude implement it.

risyachka··on Vibe code is legacy code
It’s not a “hack” when you have your stripe keys in the frontend.

It’s called a skill issue and to prevent these you have to actually spend time learning instead of vibing.

It’s literally the same as driving a car without license.

Everyone can do it - but shouldn’t.

risyachka··on Ask HN: Why is "Tea" still on the App Store after so many data breaches?
Tbh this is possible only in software. No matter what you do - epic incompetence, leak user data, doxx users, basically allow their identities to be stolen etc - zero consequences.

Kinda crazy. In any other industry they would not even allow you in the door without showing some king of understanding what you do.

You can't even sell hotdogs without food license. But in software - wild west.

risyachka··on Ask HN: What are you working on? (July 2025)
The easiest way to share development iOS (ipa) or Android (apk) builds Free to use

https://appsend.dev

risyachka··on H-1B program grew 81 percent from 2011 to 2022
You should probably do some research on how many of those ended up starting their own companies. And how many jobs they created.

After that you may want to rethink suspending new entries.

risyachka··on ChatGPT agent: bridging research and action
>> many are optimizing happy paths in their demos and hiding the true reality

Yep. This is literally what every AI company does nowadays.

risyachka··on What caused the 'baby boom'? What would it take to have another?
Everything worth doing is hard.
risyachka··on Context Rot: How increasing input tokens impacts LLM performance
Yep. The easiest way to tell someone has no experience with LLMs is if they say “RAG is dead”
risyachka··on Building a Mac app with Claude code
You learn by trying to solve the problem by yourself.

This is not about syntax but about learning how to create solutions.

When you read solution you merely memorise existing ones. You don’t learn how to come up with your own.

risyachka··on Building a Mac app with Claude code
The main issue here is you can’t acquire expertise by using llm to code for you.

So unless you have 15+ yoe better add more reps. You can always switch to llm code assist in a blink, there is no barrier to entry at all.

risyachka··on Developing with GitHub Copilot Agent Mode and MCP
>> is going with MCP, tools etc.

all these are just tools. there is nothing more to it. there is no etc.

risyachka··on Cloudflare Introduces Default Blocking of A.I. Data Scrapers
Maybe so, but I'll take Cloudflare over OpenAI and Meta every time.
risyachka··on The new skill in AI is not prompting, it's context engineering
“A month-long skill” after which it won’t be a thing anymore, like so many other.
risyachka··on Ask HN: What Are You Working On? (June 2025)
Open source Apple’s hide my email alternative

https://github.com/webmonch/hide-my-mail-cloudflare

← PreviousPage 8 of 23Next →