Hacker News: dcliu

New comment by dcliu in "DaVinci Resolve – Photo"

dcliu — Tue, 14 Apr 2026 03:14:55 +0000

DaVinci Resolve has been an incredible value. Hoping this becomes a viable contender vs Capture One and Lightroom.

New comment by dcliu in "Show HN: Cerno – CAPTCHA that targets LLM reasoning, not human biology"

dcliu — Tue, 31 Mar 2026 20:10:32 +0000

Interesting idea but it didn't accept me the first 3-4 times I tried to draw the path on laptop trackpad - going slow led to it failing before I was finished, going fast led to messy paths that were rejected as well. Could something like this work for users with different levels of dexterity?

New comment by dcliu in "E2E encrypted messaging on Instagram will no longer be supported after 8 May"

dcliu — Fri, 13 Mar 2026 14:07:24 +0000

On the other hand Messenger has moved to only supporting e2ee chats, wonder why the difference.

New comment by dcliu in "The trackball that merges pointing and 3D control"

dcliu — Sun, 08 Mar 2026 16:38:12 +0000

Creator here - yes, it's a mod of a Kensington SlimBlade Pro, mentioned on the site. Rotatrix is the hardware mod (added controller), firmware, and software stack that extracts 3DOF rotation and integrates into CAD and other apps. It's possible with firmware, but you need a trackball with dual optical sensors to do this (which very few have), and the system is patent pending. More on the architecture in the earlier Show HN: https://news.ycombinator.com/item?id=46990422

Thanks for taking a look.

Show HN: Rotatrix – Trackball mod capturing full 3-axis rotation for 3D control

dcliu — Thu, 12 Feb 2026 15:59:33 +0000

Hi HN - I've been on creative sabbatical building hardware and software projects, and this is one of them.

Rotatrix is a hardware mod of the Kensington SlimBlade Pro trackball that reads out full 3-axis ball rotation (XY plus twist) and maps it to continuous 3D control. The original device only used twist for discrete scroll gestures, which is a whole degree of freedom that can be put to more use.

For 3D control, I've found the most natural mapping is for the ball top to map to the object front, so normal trackball directions (up/down, left/right) become pitch and yaw, matching how rotation works with 2D “click and drag”, while twist rolls the object on screen. It feels natural for translation too: XY maps to familiar mouse motion, and twist corresponds to depth. Since you can twist the ball by moving a finger tangentially on the side of the ball, it feels like pushing and pulling the object.

I routinely use a SpaceMouse for 3D work, and one difference that stands out to me is that this is position control, not rate control (like a mouse vs a joystick). This gives a more direct control feel, like you're physically manipulating the object. Position control has a tradeoff: you have to re-grasp the ball to continue past your hand's range of travel. For large continuous motions, I added a rate mode; the ball's large range of travel actually gives finer rate control than a small-displacement device like a SpaceMouse puck.

How it works:

- I added a custom microcontroller inside the SlimBlade Pro alongside the stock one. It reads raw motion data from the ball's optical tracking hardware and presents its own interface to the host computer.

- Host software calibrates the raw sensor data into clean 3D rotations, then applies configurable per-app bindings and outputs via an open protocol.

- The original trackball controller is untouched except that it's no longer connected to the USB port; it still works over Bluetooth/wireless.

Current state: working prototype, configurable per-app profiles with modal bindings (hold a key to switch what the ball does), dominant axis weighting for single-axis precision. I’m building out integrations with different 3D apps and I’m looking for early users to try it and give feedback, particularly people doing CAD, 3D modeling, or geospatial work.

Happy to answer questions about the product, the UX design decisions, or 3D input in general.

Comments URL: https://news.ycombinator.com/item?id=46990422

Points: 3

# Comments: 0

New comment by dcliu in "Open Source Implementation of Apple's Private Compute Cloud"

dcliu — Thu, 06 Nov 2025 20:22:51 +0000

That's a strong claim for not looking into it at all.

From a brief glance at the white paper it looks like they are using TEE, which would mean that the root of trust is the hardware chip vendor (e.g. Intel). Then, it is possible for confidentiality guarantees to work if you can trust the vendor of the software that is running. That's the whole purpose of TEE.

New comment by dcliu in "FTC takes action against Uber for deceptive billing and cancellation practices"

dcliu — Mon, 21 Apr 2025 22:46:45 +0000

Glad to see this happen. As described in the article, I recently got signed up and charged for Uber One, despite having opened the app in weeks.

New comment by dcliu in "Twilio is laying off 17% of workforce"

dcliu — Mon, 13 Feb 2023 19:22:43 +0000

I am just learning about too, but I think the explanation goes something like this: Because the risk-free rate of return is now much higher, the net present value of the estimated future cash returns from that investment has gone down. Another way to look at it, earning a dollar today is now worth much more than earning a dollar in five years.

Just how much has it increased? If we use the 3-month Treasury bill rate as the risk-free return, it has gone from 0.05% to 4.64%. A dollar in five years used to be worth (1 / 1.0005^5) and now it is only worth (1 / 1.0464^5). That's a drop of 25%, definitely significant enough to make some speculative and longer-term investments no longer pencil out.