Hacker News: thomasikzelf

New comment by thomasikzelf in "The HTML-in-Canvas API origin trial"

thomasikzelf — Fri, 23 Oct 2026 07:01:52 +0000

This had to do with leaking privacy sensitive information such as visited links by reading back pixels from a rendered . They fixed this with a concept they call "read-back-allowed rendering"

New comment by thomasikzelf in "FLUX 3 Image"

thomasikzelf — Sat, 03 Oct 2026 09:04:39 +0000

In the context of knee bending: I was creating a simple walking animation with gemini 3 pro and could animate all the poses instead of one, left leg bend right leg straight. Right leg bend it had no problem but this one it just did not seem to understand.

New comment by thomasikzelf in "Getting out of the way: my robotics crash course"

thomasikzelf — Thu, 01 Oct 2026 12:49:59 +0000

With a car I can imagine asking questions like "am I on the left side of the lane", "am I closer then 10 meter to the next car" which feel like questions the current models could answer. I see models understanding the calibration of the robot. Maybe you could have a jev style model make decisions like "move left", "activate gripper" and do the inverse kinematics in a different system.

New comment by thomasikzelf in "Getting out of the way: my robotics crash course"

thomasikzelf — Thu, 01 Oct 2026 12:43:32 +0000

Nice work! Vision on the qwen models works very good. Jev is a really neat idea. I tried replicating jev style with qwen models using max 1 token output, which works great. How does this compare to that in terms of speed?

New comment by thomasikzelf in "Getting out of the way: my robotics crash course"

thomasikzelf — Wed, 30 Sep 2026 18:39:08 +0000

What kind of decisions though? jev does not output joint angles, and if it did it also needs to know how. It also does not take in images.

New comment by thomasikzelf in "Jeff – Jev-compatible 0.8B decision models, trained at home, ~30 ms"

thomasikzelf — Mon, 28 Sep 2026 23:55:34 +0000

Great project! How does this compare to asking qwen to reply with 1 token in terms of speed?

New comment by thomasikzelf in "Businesses with ugly AI menu redesigns"

thomasikzelf — Wed, 22 Jul 2026 15:03:29 +0000

+ design looks better then the "minimal" style original

- dish in the top left is not on the menu (why is it there)

-- why is there a malnourished cow on the menu?

++ photos of of the product so you can see what you are gonna get

--- the photos do not match the actual meal

New comment by thomasikzelf in "The unreasonable difficulty of time series forecasting"

thomasikzelf — Wed, 22 Jul 2026 12:36:58 +0000

I did some energy price forecasting and my neural network learned the morning evening patterns just from historical data, it worked great. Then I tried adding more information to this like wind and solar (and many more). None of these external factors made any difference while being very important for energy prices. It seems like the market already incorporated all the factors in the price (as they should).

The takeaway might be that historic data of market data might often be enough to make a reasonable prediction. Only external data that nobody else has (used) can make your prediction better then the market.

Making a prediction with same accuracy as the market: easy Making a prediction with more accuracy then the market: very hard

New comment by thomasikzelf in "Show HN: A new kind of FPS aim trainer"

thomasikzelf — Wed, 22 Jul 2026 10:01:56 +0000

Cool app! I also had some ideas a while ago that I think the current breed of aim trainers are still missing.

In-game practice like the Range, DM, and Team DM still feels like the gold standard to me. I generally only use aim trainers when they offer a noticeably higher return per minute, rather than just teaching me how to get good at aim trainer specific mechanics.

Here are a couple of things I think an aim trainer should have:

- Dynamic difficulty: I feel it works best when trying to keep accuracy relatively high, but fluctuating the target accuracy threshold over time and adjusting difficulty dynamically to match it. That keeps you challenged without breaking form.

- Dynamic scenarios: Instead of picking scenarios manually, I think a trainer should ideally detect what you suck at and launch those directly.

- Habit coaching: It would be awesome if a coach flagged bad habits in real time—like clicking too fast (low accuracy), doing a slow tracking movement instead of a sharp flick (mircoflicks), or using tracking when click timing is needed.

- Biomechanics: Finding the exact "perfect" sensitivity feels less important to me than ensuring your arm and wrist are working together properly, altough this might depend on your aiming style.

- Simplicity: A lot of trainers overwhelm you with numbers and menus. I image a simple "click here to start and follow feedback" workflow with nothing else for the user to learn. Seeing a bunch of numbers is less important to me then just getting better (maybe with a final score).

Valorant-specific focus: Trainers often feel too generic for tactical shooters like Valorant, where angle holding, target reading, and micro-corrections dominate.

I built a small aim trainer to experiment with these ideas myself: https://mousecontrol-thomaswelter-fceff2dbe456668c6bedef32db... (the aim trainer is very basic, only made for myself, fixed sens etc).

- In this tool, difficulty scales dynamically while playing (the parameter in parentheses changes on the fly):

- Angle Hold (enemy speed increases): Practicing holding angles against multiple targets. Holding angles is crucial in Valorant (just look at how Primmie plays).

- Follow (standstill window decreases): Forces active mouse correction and reading the target before shooting. I tend to click too fast, so this forces me to slow down and micro-correct.

- Jiggle (jiggle speed increases): Targets perform tight jiggles. I find players who spam small jiggles really hard to hit, so this directly targets that weakness.

- Micro Flicks (snap distance increases): Micro flicks are easy enough to train in the Range, so I mostly use this mode when I am away from my gaming PC.

- Static (target size decreases): Standard static clicking, but with dynamic scaling as targets shrink.

- Strafe (movement speed increases): Moving targets with limited hit windows. When watching pro DM VODs, most kills are on stationary targets, but hitting moving players is still vital and tricky to isolate in-game.

I am still not very good at Valorant but the aim train ideas are interesting. Let me know what you think!

New comment by thomasikzelf in "Hyundai buys Boston Dynamics"

thomasikzelf — Fri, 19 Jun 2026 19:10:36 +0000

I have visited factories for work and my experience is the same. There is so much stuff that could easily be automated but is not because it is too expansive for too little value to make a custom one off machine. The big high volume things will be automated but these machines will have 90% success rate and lot's of stuff that needs to be done by hand. You can search for factory tours on youtube to get an idea. Here are two videos, an Amazon warehouse and a Tesla factory. the big heavy stuff is automated but lot's of work is still done by hand. https://www.youtube.com/shorts/-R6cBkza17k https://www.youtube.com/watch?v=45slYC99uUg

New comment by thomasikzelf in "Building a robotics research setup that lives next to my desk"

thomasikzelf — Fri, 19 Jun 2026 12:38:52 +0000

It looks like one stole the design from the other, I don't know which one, haha.

New comment by thomasikzelf in "Building a robotics research setup that lives next to my desk"

thomasikzelf — Fri, 19 Jun 2026 06:58:50 +0000

Nice, I will be following your posts! I just bought a robot arm myself, the seeed studio B601DM (€1500 6+1 axis), it works great and is open source hardware as well and a bit more solid then the so101. I also opted to not use ros, I don't want to give up control by putting another framework in between. Is your plan to see whats possible right now or do you also have ideas on how to improve sota?

New comment by thomasikzelf in "CSS-Native Parallax Effect"

thomasikzelf — Tue, 02 Jun 2026 12:26:37 +0000

firefox android does not support CSS animation-timeline, and firefox desktop needs layout.css.scroll-driven-animations.enabled. This probably should not be used for any critical features.

New comment by thomasikzelf in "CSS-Native Parallax Effect"

thomasikzelf — Tue, 02 Jun 2026 11:26:53 +0000

You can make some really cool stuff with css scroll animations. I used SVG paths with a scroll animated dash offset to draw an image while scrolling. Zero javascript, it feels so smooth. https://thomaswelter.nl (the background)

New comment by thomasikzelf in "The three pillars of JavaScript bloat"

thomasikzelf — Sun, 22 Mar 2026 09:49:51 +0000

The declarative vs imperative example is strange here. Why is the imperative example so convoluted? This is what one could write in js:

  badge.textContent = count > 99? '99+' : count
  badge.classList.toggle('show', count > 0)
  paper.classList.toggle('show', count > 0)
  fire.classList.toggle('show', count > 99)

The declarative example also misses the 99+ case. I don't think this example describes the difference between imperative and declarative well.

New comment by thomasikzelf in "I was banned from Claude for scaffolding a Claude.md file?"

thomasikzelf — Thu, 22 Jan 2026 23:20:22 +0000

I was also banned from claude. I created an account and created a single prompt: "Hello, how are you?". After that I was banned. An automated system flagged me as doing something against the ToS.

New comment by thomasikzelf in "The Easiest Way to Build a Type Checker"

thomasikzelf — Sun, 30 Nov 2025 17:51:32 +0000

I recently implemented a Hindley Milner type checker. The algorithm itself is not necessarily super difficult (once you get your head around it ofcourse) but the way it is explained is. It seems like HM is mostly explained by people with a mathematics background that already know the proper notation. I wonder how much overlap there is between people that know the notation and do not know how HM works, probably not much.

Anyway nice demo. Bi-directional is already quite powerful!

New comment by thomasikzelf in "`satisfies` is my favorite TypeScript keyword (2024)"

thomasikzelf — Sun, 23 Nov 2025 15:24:16 +0000

Really, that is surprising to hear. There are a couple of differences but most of the syntax looks the same to me, what part do you find alien?

The reskill problem is of similiar difficulty with learning a new framework I think. Especially because the language is rather simple compared to typescript (which is also its strength).

I do understand it is an uphill battle. The whole nobody get's fired for choosing IBM thing. The language is still unproven in the general perception. I do think that when it comes to libraries and frameworks I see a lot of developers choose new unproven stuff, more then they do languages.

New comment by thomasikzelf in "`satisfies` is my favorite TypeScript keyword (2024)"

thomasikzelf — Sun, 23 Nov 2025 11:35:46 +0000

Have you seen ReScript? Of course it is not as popular as typescript but it improves on all the bad parts of typescript. You'll get sound types with the pain points of javascript stripped out. Because it compiles to readable javascript you are still in the npm ecosystem.

You don't have to rewrite your whole codebase to start using it. It grows horizontally (you add typed files along the way) compared to typescript which grows vertically (you enable it with Any types).

The point is that we don't have to move back to plain js. We have learned a lot since typescript was created and I think the time has come to slowly move to a better language (and ReScript feels the most like Javascript in that regard).

New comment by thomasikzelf in "`satisfies` is my favorite TypeScript keyword (2024)"

thomasikzelf — Sun, 23 Nov 2025 09:02:08 +0000

If Typescript is javascript with types bolted on, Rescript is javascript with types the way it should have been. Sound types with low complexity. https://rescript-lang.org/