HNHacker News
TopNewBestAskShowJobs

risyachka

1,090 karma · joined November 7, 2017

submissionscomments
risyachka··on Claude Fable is relentlessly proactive
>> Getting the best results out of these models requires skill, experience, intuition, and domain expertise.

domain expertise has nothing to do with llms. On the contrary, to have it you need to avoid llms.

>>you risk prompting Claude Fable 5 like it's GPT-4o

Thats fine because when GPT came out you had to treat it like a baby, GPT2 and around that time "Prompt engineering" was a thing.

Now its all dead.

After opus 4.8 all you have to do is say "fix it" or add /plan. All that time spend on learning previous models is time wasted.

And in a year or two with developed harness you will be out of the loop, errors are incoming - llm fixes them or adds new features based on some transcripts etc.

Even if model development stops now - there is nothing to learn really. Sure you may need to adjust prompt style a bit. You will do it naturally just like when you communicate with a new person. There is no "knowledge" to it, it is very smart.

risyachka··on Claude Fable is relentlessly proactive
>> he would have learned nothing about the AI

there is absolutely zero value in spending time to learn about new models as in few months new model will be out and whatever you learned about the current one will be useless.

Also with models getting better and better you have to know less and less to achieve same results.

risyachka··on I design with Claude more than Figma now
This. LLMs helping implement many ideas is a curse and the results speak for themselves.

As a non-designer I want to see a thought through 1-2 ideas, not 10 ideas you coded with llm and now the burden of thinking which one is better and why is for some reason is on me.

risyachka··on I design with Claude more than Figma now
>> It’s more effective to throw everything in using LLMs, then using human judgment to sculpt away.

Not really. Maybe its effective from designer side but from other side that has to review all those ideas it is exhausting and counterproductive. Especially when it is in code.

You really don't need a real prototype for most things. Simple visual is more than enough to present an idea and explain etc. In fact, it is often much better as you can see all flows in canvas at once.

With interactive prototype to see all flows of a feature you have to go through each one, which again is very uneffective.

risyachka··on Changing how we develop Ladybird
This is direct result of AI as you can see in many other public repos.

before AI like 1 in 1000 would spend their time fixing something they had no idea about and even then considering how much time you spent and how few of those happened it made sense to review/talk about it.

now every "dev" with claude submits prs having absolutely no idea what they are even doing. most of them would not even be able to create PR without AI in the first place.

and on top of that add slop bots that "fix" issues in the loop and create hundreds of PRs daily

risyachka··on The dangerous delusion of modern warfare
> Both are wars instigated by the leaders of great powers

having nukes and a lot of empty land does not make russia "great power", and their war results match this.

risyachka··on Cloudflare Flagship
they are quickly turning into a slop shop

instead of polishing their existing products (and most of them do require a lot of work) they jump into any other niche someone thought was a good idea. My guess is that with ai being able to prototype things quickly they just started doing everything that is even a bit relevant.

which won't end well

risyachka··on Migrating from Go to Rust
In order to use C you need to actually understand it, also toolchain is more complex etc. Which makes it a no go for 99%.

Rust is so safe that anyone can vibe it without any idea what is going on there. Which is basically what is happening here.

And why rust is more used than go for vibecoding? Mostly because of hype and performance gains which 99.9% of projects do not need.

risyachka··on Microsoft open-sources "the earliest DOS source code discovered to date"
Except the competition was essentially non existent and no one would copy your product with llm in a day
risyachka··on Green card seekers must leave U.S. to apply, Trump administration says
>> so that even legal immigration is restricted or difficult

like it was simple and easy before that. Now it becomes borderline impossible

risyachka··on Open source Kanban desktop app that runs parallel agents on every card
No one reads code that results from this. Those who say otherwise either lie or are very bad developers which is essentially the same as not reading that code.
risyachka··on Flipper One – we need your help
just fyi its also russian
risyachka··on Learnings from 100K lines of Rust with AI (2025)
I may have missed it but are those tests written by person or generated? Otherwise how do you know they even test anything (like actually test, not appear to test)
risyachka··on OpenClaw Creator Spent $1.3M on OpenAI Tokens in 30 Days
Thats the single reason it is faster. Just pushing to prod whatever.

All projects can become fast if they drop guardrails.

This does not correlate with productivity increase

risyachka··on Bun Rust rewrite: "codebase fails basic miri checks, allows for UB in safe rust"
This is because AI is most appealing to average and below average developers and users because it makes them feel like they can finally do something.
risyachka··on Bun Rust rewrite: "codebase fails basic miri checks, allows for UB in safe rust"
At this point bun is a free adspace for Anthropic and Claude. This is the only thing that would explain the rewrite
risyachka··on I believe there are entire companies right now under AI psychosis
Humans can already have 4 hour work week without productivity loss.

But I only see mass layoffs and those who are working - are working longer and harder then before.

risyachka··on Show HN: Building a web server in assembly to give my life (a lack of) meaning
The analogy is AI music produced without effort in seconds, not recorded music.
risyachka··on Show HN: Building a web server in assembly to give my life (a lack of) meaning
llms are nothing like hammers or other tools.

They are factories that product goods on a whim. There is nothing to compare them to as we never had anything like that. This is not industrial revolution this is obliteration of work at its core.

risyachka··on Show HN: Building a web server in assembly to give my life (a lack of) meaning
the biggest issue with llms is that they make those who have no idea what they are doing seem like they know something.

>> without the limitation of your ability to hand code it.

yeah its nice though this in 100% of cases results of software of even lower quality we had before.

so hard to tell where is the win here. the fact that you can generate some code does not make it a win, just a curious fact.

risyachka··on AlphaEvolve: Gemini-powered coding agent scaling impact across fields
In coding the ambiguity is very, very limited and constrained compared to any non dev job that involves any decision making
risyachka··on Vibe coding and agentic engineering are getting closer than I'd like
This is so on point that I want to cry.
risyachka··on Nuclear receptor 4A1 linked to health effects of coffee: study
Right? All high quality coffee makers use a proper method so there is absolutely zero downside in decaf. Just make sure to check which method they use (all big ones state it on their website or else)
risyachka··on Uber torches 2026 AI budget on Claude Code in four months
>> Again, not necessarily bad in itself,

yeah, it is bad. Human brain is not able to properly assess this amount of changes. To understand even a small change you need a lot of capacity. To understand thousands of lines - impossible.

This is pure slop pouring into prod and we can see more and more consequences of this in all big corps's products - things start to break more and more exponentially faster.

risyachka··on Claude Opus 4.7 costs 20–30% more per session
How does this solve the issue? 4.6 will be disabled after one or more release like any other legacy model.
risyachka··on Android CLI: Build Android apps 3x faster using any agent
Frontend, backend, workflow engines, payments, idk all of it.

Like you can make a cross-platform react native app with expo in a day without any AI (a proper app, not a boilerplate). Same with many web apps.

Shadcn with themes and components, tailwind, temporal workflows, etc etc. The complexity of making apps was solved years ago.

risyachka··on Artifacts: Versioned storage that speaks Git
You can always have a dynamic pool based on your load of ready to use docker containers, then it takes like 15 seconds and is faster than basically any sandbox provider.
risyachka··on Android CLI: Build Android apps 3x faster using any agent
The tooling in 2026 is so easy you can do almost anything without AI very very quickly.
risyachka··on Codex for almost everything
Its not like Claude is pioneering those. All that was done prior to all of them by some random startup.
risyachka··on Building a CLI for all of Cloudflare
> Tell us your hopes and dreams for a Cloudflare-wide CLI

Please call it flare.

← PreviousPage 3 of 23Next →