HNHacker News
TopNewBestAskShowJobs

LatencyKills

612 karma · joined January 14, 2014

Senior engineer on the Visual Studio team at MS and the Xcode team at Apple. If you're old, like me, you might also remember tools like BoundsChecker and SoftICE which I had a blast working on.

I love making tools for developers.

Feel free to reach out: lookingfortheoffswitch@gmail.com

submissionscomments
LatencyKills··on Show HN: Auto-identity-remove – Automated data broker opt-out runner for macOS
Why not just comment out the macNotify() calls in watcher.js and then run it periodically? There are also a few calls to send iMessages that you should remove.
LatencyKills··on It's set up, not setup: Scraping GitHub for grammar errors
I was an engineer at MS in the 90s. We've been using the term "setup" incorrectly since the beginning of time. It was "setup.exe", not "set up.exe" for an obvious reason.

You are conflating grammar with canonical usage. Should people use the correct form? Sure. Does it matter in the least? No.

LatencyKills··on Show HN: Liquid Radius – Unifies macOS Tahoe's inconsistent window corners
> Liquid Radius doesn't touch /System or modify the SSV

I'm not doubting you, but by requiring users to disable SIP you are effectively opening their entire system to SSV attacks. 99% of "regular" users won't understand that tradeoff.

Rounded corners are not worth disabling the operating system's major security protections, especially in today's crazy cyber environment.

yabai [0] is a good example of how this type of tool can be shared. I would never install a closed-source app that requires me to disable SIP without being able to build it myself.

> are you aware of any system wide visual tweak like this that can be done without injection

There used to be several ways to do it (e.g. userdefaults) but AFAIK that is no longer possible. As I'm sure you already know, there are ways to do this to individual apps which would probably be a better use-case.

[0]: https://github.com/asmvik/yabai

LatencyKills··on Show HN: Liquid Radius – Unifies macOS Tahoe's inconsistent window corners
As an ex-Apple engineer I have a very good idea how you implemented this and why you require SIP deactivation.

With SIP disabled, malicious tools can directly patch binaries, inject shims, or drop payloads into system locations. Also, SIP works with the Signed System Volume (SSV), which cryptographically seals the system volume and enforces integrity checks.

Any app that tells a normal end‑user to “disable SIP to use me” is effectively asking them to defeat a system‑wide safety net and trust the developer absolutely.

Using this tool would be be like removing every lock in your house because you don't like the shape of the roof.

If I wanted to create malware, I'd wrap it in something like this project which ostensibly provides a service. That's far easier than trying to hack in normally.

LatencyKills··on Mouse Pointer as a Mere Mortal
> The important point is that applications shouldn't be able to arbitrarily move the mouse in situations when it's unexpected.

That is quite a different statement from "It should be impossible." What should be impossible is for the OS to prevent this type of usage when it is clearly useful. Beyond accessibility, I use these features to automate testing of native macOS GUI apps.

LatencyKills··on Mouse Pointer as a Mere Mortal
> But if it's windowed then it should be impossible.

I worked on several apps for the visually impaired that automatically move the mouse cursor to different UI elements in the front-most application, regardless of the window state. It’s a good reminder that “impossible” often just means “I haven’t accounted for that use case yet.”

LatencyKills··on Can anyone point me to a ShowHN first post policy documentation?
You can send an email to hn@ycombinator.com. They tend to respond promptly.
LatencyKills··on Ask HN: How to Sell as a Software Engineer?
> I've reached the conclusion that software quality and even price matters much less than distribution, especially now.

I’m skeptical of that claim. Even if it's trivial to build something that “works,” we still don’t know how users will make purchasing decisions in that world. AI coding won't change the fact that customer retention still depends upon software quality and service.

AI or not, the fundamentals haven’t changed much: you need to define a target segment before building, and you still need a way to reach those users.

For example, I build macOS utilities for power users, and distribution channels like Setapp [0] (no affiliation) work well because they pre-select for that audience. Users there are already inclined toward the kinds of tools I make.

WRT to software quality, my experience is the opposite. Users care about thoughtful features, solid UI/UX, and reliability. If anything, an increasing supply should make those differentiators more important, not less.

[0]: https://setapp.com

LatencyKills··on How to disable Firefox's new emoji picker
Neither MS nor Apple has the best track record of building full-featured utility tools (which is why there are thousands of incredibly useful 3rd-party utility applications). I wish end-users would understand that giving them a choice is more important than forcing them to use only OS-provided functionality.

You aren't required to use anything that you don't like. The fact that you are complaining about a non-problem is odd.

Lastly, incremental improvements generally come from outside the OS. If FF ends up adding something novel/interesting to their emoji picker, the major OS's will probably adopt it at some point. That would never happen in your world, where only the biggest companies get to innovate.

LatencyKills··on Tar Files Created on macOS Display Errors When Extracting on Linux (2024)
Ex-Apple engineer here. This is, for better or worse, just the way Apple approaches this type of problem. From Apple's perspective, this is the way to preserve Finder / Gatekeeper / metadata semantics. It avoids silent data loss when round-tripping archives between Macs. This behavior also maintains consistency with copyfile(3) (as well as the Archive Utility behavior).

Apple treats tar less like “portable Unix interchange” and more like “archive this filesystem object faithfully.” That is very Apple, and very libarchive. ;-)

This is probably going to get worse (as Apple continues to add macOS-specific metadata), so your workaround is very helpful.

I haven't tested it in a while, but at one point, setting the COPYFILE_DISABLE=1 env variable would disable the inclusion of macOS-specific metadata.

LatencyKills··on How to disable Firefox's new emoji picker
I understand that, but as someone who was an engineer at both Microsoft and Apple, I see absolutely no problem with 3rd-parties adding useful features.

As I said, as long as 3rd parties aren't overriding the built-in functionality by default (e.g., using the same keyboard shortcuts), there's no problem. 99% of FF users will probably never even know it has its own picker.

LatencyKills··on How to Disable Firefox's New Emoji Picker
Allow me to rephrase: If the OS doesn’t provide X, the end user should have no alternative for X.

I use both macOS and Pop!OS. The latter doesn't include an emoji picker by default, so I'm precluded from using a 3rd-party picker?

There is no issue as long as a 3rd-party app doesn't override built-in functionality. If you don't want it, it is easy to disable.

LatencyKills··on Show HN: Drive any macOS app in the background without stealing the cursor
Ex-Apple engineer here. I really like your implementation. A few years ago I built a similar tool to help me automate the testing of some of my native macOS apps. Being able to run multiple UI automation tests simultaneously was the big win in my case.

My only criticism is enabling telemetry by default. I'm a fan of having people opt-in.

LatencyKills··on DeepSeek Slashes Fees for New AI Model
> DeepSeek-V4-Pro API is 75% OFF until May 5th

Since the API price reduction is temporary, I wonder if this is intended to counter OpenAI's temporary limit boost [0].

[0]: https://venturebeat.com/orchestration/openai-introduces-chat...

LatencyKills··on OpenAI Privacy Filter
> If you have the redacted and unredacted versions, then you can diff them; that seems unsurprising?

I'm suggesting that a model designed for high-accuracy redaction can also be used to find all PII in unredacted text. For example, if I don't already know how to find PII (e.g., regex, NLP, etc.) I can use OpenAI's Privacy Filter model to do the work for me.

And because each span has a type (PRIVATE_NAME, etc.) I don't even need to do any work to find only the specific information I am looking for; something that simple diffing wouldn't do.

I'm not saying it's an issue, I just think it is interesting that a tool designed to protect PII can also be used to find it with minimal effort. And it looks like someone already implemented it: https://github.com/chiefautism/privacy-parser.

LatencyKills··on OpenAI Privacy Filter
Couldn't this be used to locate private data in unstructured text without having to rely on other means of PII detection?

1. Pass the raw text through the filter to obtain the spans.

2. Map all the spans back to the original text.

Now you have all the PII information.

LatencyKills··on An update on recent Claude Code quality reports
That is exactly what I do. The bash script runs, determines that a code file was changed, and then is supposed to prevent Claude from stopping until the tests are run.

Claude is periodically refusing to run those tests. That never happened prior to 4.7.

LatencyKills··on An update on recent Claude Code quality reports
I have a set of stop hook scripts that I use to force Claude to run tests whenever it makes a code change. Since 4.7 dropped, Claude still executes the scripts, but will periodically ignore the rules. If I ask why, I get a "I didn't think it was necessary" response.
LatencyKills··on Work with the garage door up (2024)
I was also a corporate stooge (at both Apple and Microsoft). One of the biggest differences in culture was exactly this point. At Apple, it was encouraged to share your successes and failures in real time across teams. At Microsoft, there was so much infighting and competition that you tended to share nothing until you were ready to release. For example, someone on the Windows team wouldn't want to give away a cool solution that could be "stolen" by the dev tools team. This resulted in walled gardens that were protected at all costs.
LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
> You cannot use a ChatGPT subscription with a CLI tool

Okay, I'm done here. You obviously have no idea how this works (I have a ChatGPT subscription that I use with Codex).

I know you're new to HN, but when someone says: I've literally built subscription tools at both Microsoft and Apple for over 25 years, you might want to stop and reconsider if you might be missing something. You are.

/blocked

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
> I still don't know why creating an API key and loading it up with $200 every month is an unfavorable option.

If I pay $200/month for a Max subscription, I can access ALL of Anthropic's tools (CC, Design, Cowork, etc.).

If I pay $200/month for API access, then the only thing I can use is the API.

You don't see how ridiculous that is? No other SOTA model company has these restrictions, which is why Anthropic keeps losing subscribers.

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
And I was referring to the all of the historical flip-flopping. This new flipping is just proving the point.

Of course, you're simply being pedantic. Everyone knows why they are making this change (which is more important than your silly take on what constitutes flip-flopping).

The point: Anthropic is losing subscribers because it has no idea what it actually wants to be.

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
> Keep in mind this is hearsay

How, exactly, is that not saying something about the announcement?

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
Given the lack of clear communication and the fact that their primary competitor openly supports the use of bespoke harnesses, I highly doubt this is an incorrect announcement.

Anthropic is destroying goodwill that is hard-won in this space. At the end of the day, people just need to do their work in a way that makes sense for them. In my case (someone who has been building ML/AI tools for 25 years @ MS & Apple), I have much better results using my bespoke harness. If I'm paying $200/month for compute, I should be able to use it in a way that works for me. Given the push back, I'm not alone.

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
The poor communication and flip-flopping are what concern me.

How can I buy into an ecosystem that might disallow one of my main workflows? I currently use several hook scripts to route specific work to different models. Will they disallow that at some point? We don't know because they can't get their story straight.

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
> I find it a little bizarre that people have this expectation.

Well, enough people complained that Anthropic reversed their stance. Additionally, their primary competitor doesn't have any compute restrictions, which should help clarify why this decision was made.

As someone who has been building ML/AI tools (@ MS & Apple) for almost 25 years, I can say that much of the value of the underlying model comes from the harness. Why shouldn't I be able to use the exact same compute with my own bespoke harness when the compute cost is the same?

The Claude Code team continues to push out half-baked features that literally hamper my ability to use their tools.

If I'm paying $200/month for compute, I should be able to use it however I like.

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
If that was the argument, why did they reverse it?

Anthropic is constantly destroying goodwill and now seems to be in panic mode.

LatencyKills··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
Many people likely objected to the original restriction because it seemed as though Anthropic was trying to impede the development of competing tools.

If I'm paying for compute, why should it matter whether I use Anthropic's harness (e.g., Claude Code) or a 3rd-party harness?

LatencyKills··on Ask HN: Is Claude Getting Worse?
I've been an engineer for almost 30 years (@ MS & Apple). I've been using Claude to perform code reviews for several of my macOS apps (that I wrote without AI).

A few weeks ago, its performance was impressive and helpful.

In the last week, it has been unusable. It is now getting confused, suggests architecture changes that make absolutely no sense, and has even started ignoring my stop hooks (and then arguing with me that they "aren't necessary").

LatencyKills··on Bill Gates to testify before House Oversight in Epstein probe
I'm offended by the fact that, per emails and Melinda's comments, he asked Epstein for STD drugs that he could give Melinda without her knowledge.
← PreviousPage 3 of 7Next →