HNHacker News
TopNewBestAskShowJobs

senko

10,074 karma · joined October 19, 2008

Ex-Pythagora (YC W24), built & sold A Web Whiteboard (acq'd by Miro) and two earlier startups. Runs GoodCode.io (dev agency). Currently doing AI-related consulting.

Interests: AI, startups, open source.

Twitter: @senkorasic

Bluesky: @senko.net

Work: https://vibescale.eu

Personal: https://senko.net & https://blog.senko.net

GitHub: http://github.com/senko

Mail: senko@senko.net

submissionscomments
senko··on Stateless MCP has recaptured my interest
To everyone saying "this is just REST": no. It's just RPC (in case "jsonrpc" in the payload didn't give it away).

If we're talking protocols, it's useful to know the distinction.

There have been many flame wars around REST vs RPC and whether most "RESTful" implementations out there are actually RPC with a funny nose.

senko··on Pi's Minimalism Is Its Advantage
Meta: this reads like a corporate article, info-selling a solution to enterprises. The style is a far cry from anything Mario or Armin usually write.

First indication of business strategy around Pi's reverse acquisition of Earendil, perhaps?

senko··on “Gravity is worth asking about”
Some people, when they want to hide an option, think: "I know, I'll use an option."

Well, now they have two options.

senko··on Waymo in Dallas
> Tesla Cybercab has about three or four times the efficiency and like a fifth of the price by having smaller, two person vehicles

I read this as Tesla Cybertruck and was mightily confused for a minute.

Looking it up in Wikipedia I see it has no driver support (steering, pedals, etc). Is that even road-legal?

senko··on The AI Productivity Gap
Of course you can beat it (in this context). You can cheat (by shifting the work around).

Started to write up a comment, got a bit wordy, turned into a blog post: https://senkorasic.com/articles/ai-amdahls-law

senko··on AI's debt binge can't last, hidden borrowing reaches $1.65T
> What would this do to the insurance or pension firms who hold this debt?

If the debt is serviced, nothing.

senko··on AI's debt binge can't last, hidden borrowing reaches $1.65T
So, what happens when it stops in this particular instance?

The article is about hyperscalers which are massively profitable irrespectively of AI. If the lending stops and this leads to paused or cancelled infra projects, but these companies are still able to service their existing debts, how does this inevitably lead to a collapse?

Hoping for a more nuanced analysis than “la la la we all know this time isn’t different”. Every time is different, and it’s instructive to know how different each time is.

senko··on Bending Spoons makes first post-IPO acquisition with $1.3B Airtable deal
I've been researching the alternatives as I've been really.annoyed by the Meetup.com enshittification.

Any other platforms besides Luma you hear about? I know some lean on WhatsApp, some tried to use FB or even LinkedIn events but these are really made for different use cases.

senko··on Software for One
Now, see, when AI does something like this, we call that "hallucination".

Thomas explained what he means in a sibling comment an hour before yours.

senko··on The End of an Era
Wow, what an epic misread :)
senko··on Azulejo
Azul[0] is a great board game inspired by this style.

Easy to start, can get quite complex, heartily recommended!

[0] https://boardgamegeek.com/boardgame/230802/azul

senko··on Superlogical
I still don't understand why tmux authors wasted their time on tmux, tho, when we already had perfectly servicable GNU screen.

Why reinvent the wheel?

/s

senko··on Open Source is not a business model (2016)
Article author here, thanks for sharing!

I mainly wrote this as a reflection on many accounts of OSS maintainers burning out.

Much has changed since 2016, but I believe the point still stands. The arrival of AI based coding (and Pull Request floods), and the tendency of many VC backed startups to "open source" their components has made the picture more complicated.

Last year I wrote a sort of a followup, geared towards founders and businesses, on how to use open source as a part of a business strategy: https://senkorasic.com/articles/open-source-strategy

senko··on Mondragon Corporation – a federation of co-operatives
They might be rare, but they do exist: https://en.wikipedia.org/wiki/Igalia
senko··on Kimi Linear: An Expressive, Efficient Attention Architecture
Not an expert, but looks like they did a lot more work on the RL part (9 expert models, full sandbox access for agentic tasks, etc)?
senko··on Kimi Linear: An Expressive, Efficient Attention Architecture (2025)
Old but relevant: if you read the recently-released Kimi K3 paper[0], you'll see that it's heavily based on Kimi Linear discussed here, scaling it up and adding a bunch more things (like native vision and RL improvements).

[0] https://arxiv.org/abs/2607.24653

senko··on Elevated Errors for Opus 5
You mean like GitHub?
senko··on How to Sell a Kettle
It's hilarious.

The Economist often has satire columns, mostly related to office or biz absurdities.

senko··on Writing by hand is good for your brain
> Being left handed is fine

As a lefty, I would rather never write another word then twist my hand the way he's doing (above the current line) to avoid the smudges.

senko··on How to Sell a Kettle
> The Ketgpt takes the kettle into the agentic era. At the front door and fancy a cup of tea? Just woken up and desperate for your first coffee? Get out your phone, fire up the Ketgpt app, enter your login details and tell your kettle to start boiling. That’s not all. You can ask your kettle how the boiling process is going in text or voice mode, and it will reply. It will even send a notification to your phone when the boiling process is complete. You set the task. The Ketgpt will take it from there, entirely autonomously. (Token costs not included.)

I have seen funded AI startups that made less sense.

senko··on Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA
FWIW my observation was about the models I tested, but I can see how it could be taken as a general statement.
senko··on Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA
For the web app task I mentioned:

* Kimi K3: 9532k input (9172k cached), 114k output - cost $5.5

* Qwen 3.8 Max: 18020k input (17836k cached), 114k output - cost $6.3

* Fable: ~14m input (all cached??), 196k output - cost $30

Correction on my earlier post, Kimi was through Pi, not Kimi Code. For Qwen I used Qwen Code and for Fable I used Claude Code.

Not sure wtf is going on with the Fable stats (a lot tokens, virtuall all of them were cached - I guess heavy system prompt?) but both claude code stats and ccusage tool output match.

senko··on Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA
Having tested K3, Qwen 3.8 max preview, Fable and Sol for the past few days, I partially agree. Don’t trust the benchmarks, and the Chinese models really are slow and token-inefficient.

However they do seem very close to SOTA : I’d say roughly equal to previous gen (Opus 4.8, GPT 5.5). It’s yet another silly benchmark, but compare them here: https://senko.net/vibecode-bench/

I also had K3, Qwen3.8 and Fable (using Kimi Code, Qwen Code and Claude Code harnesses respectively, and the official APIs) create a simple but far from trivial web app (zero shot, from a detailed spec). In user testing all three results looked/behaved more or less the same. I had Sol (via codex) do code reviews on all three and it concluded all three were solid, with some room for improvement. Fable was slightly ahead of the pack.

In my own work I still prefer Opus 4.8 (until Antrophic come to their senses and allow 100% Fable usage on Max plans) and Sol, but if I had to find an alternative, I could live with both K3 and Qwen3.8 just fine.

senko··on GPT-5.6
Here it is: https://senko.net/vibecode-bench/2026/rts-gpt-5.6-sol.html

Clearly much better than the Terra version. I'd say its on par with Fable, and the observed differencies are more due to random luck and open-ended prompt, rather than model capability. (Edit: after some more testing, perhaps not on par - somewhere between Opus and Fable, is a better description).

Fable did better pathfinding and has more terrain variety, visually the map looks better, especially soft edges of the fog of war. And the enemies.

Sol took more care with tiny ux details, added help, and more building varieties.

senko··on GPT-5.6
Here's the prompt I used:

> Create a simple but functional real time strategy (RTS) game similar to old WarCraft, StarCraft or Command & Conquer games. The player should be able to build buildings, create units, gather resources and should uncover the whole map. No AI or multiplayer needed. Use simple but nice-looking graphics. No sound. Implement everything in HTML/CSS/JS, everything in a single file (you can use 3rd-party js or css libraries/frameworks via CDN).

senko··on GPT-5.6
I'd say Fable 5: https://senko.net/vibecode-bench/2026/rts-fable-5.html

It even has enemies! (I'm not too mad about it not following my instructions because it can be fun to play :) And I generated that from Claude Code on my phone.

Sonnet 5 also produced a pretty nice version. You can see all of them here: https://senko.net/vibecode-bench/

senko··on GPT-5.6
Well, it is a silly test, not a scientific benchmark.

However, I would say it is a measure (not the measure). If you look at the entries, there's a lot of variation - definitely not something they memorized outright.

And the test itself is deceptively simple. You need to do canvas rendering, there's pathfinding, command queueing, terrain generation, etc. There are some subtle click handler bugs (various LLMs often stumble on those). And I ask the model to do it all in one file, further increasing the complexity of the task.

And the result is something that you can instantly evaluate. And if the result is any good, even play! So yeah, I think it's a fair test.

I'm sure it'll get saturated at some point. Actually I started with Minesweeper and switched to RTS last December, because Minesweeper was being saturated. I'm expecting (hoping?) the RTS test will last until the end of this year...

senko··on Buried Apple feature turns an iPhone into the perfect kids' dumb phone
You guys had filesystems??
senko··on GPT-5.6
I love testing the new models by asking them to code a toy RTS game. Here's what Terra did: https://senko.net/vibecode-bench/2026/rts-gpt-5.6-terra.html (one try, in codex app, xhigh effort)

Comparing this to other models, I find it similar to GPT-5.5 and a bit behind Sonnet 5. You can see how other models fared here: https://senko.net/vibecode-bench/ (you can also fetch the prompt and the the 5.6 Terra resulting code on from that page).

I don't have access to Sol yet (on a Plus sub, which should get it according to what I've read), so can't do the more interesting test. I'll update the above page as soon as I get access - hopefully soon.

senko··on Bonnie Tyler, singer of Total Eclipse of the Heart, dies aged 75
'fraid so, the odds have risen too high.

We might need a white night on a fiery steed. One can dream.

← PreviousPage 3 of 34Next →