garrison
Joined 16 years ago
Karma
215
Submissions
25
Member since
May 23
Recent submissions
Comment · 1 year ago
This reminds me of an earlier blog post by the same author: https://web.archive.org/web/20121111234839/http://www.altdev...
Comment · 2 years ago
Comment · 2 years ago
It also inherently takes at least 5 years from when a study starts to show that progression-free survival is at least 5 years.
Comment · 2 years ago
Looks neat. A screenshot or video would be much appreciated.
Comment · 2 years ago
If OpenAI remains a 501(c)(3) charity, then any employee of Microsoft on the board will have a fiduciary duty to advance the mission of the charity, rather than the business needs of Microsoft. There are obvious conflicts of interest here. I don't expect the IRS to be a fan of this arrangement.
Comment · 3 years ago
My recent solution to this problem -- for an entire static site -- was to use HTTP Basic authentication with CloudFlare Pages: https://github.com/garrison/cloudflare-pages-shared-password
Comment · 3 years ago
The final point, involving labels and references, could be improved by mentioning cleveref: https://ctan.org/pkg/cleveref?lang=en
Comment · 4 years ago
Similar correctness issues are a big part of the reason that, several years ago, I submitted a series of pull requests to Julia so that its entire test suite would run without memory errors under Valgrind, save for a few that either (i) we understood and wrote suppressions for, or (ii) we did not understand and had open issues for. Unfortunately, no one ever integrated Valgrind into the CI system, so the test suite no longer fully runs under it, last time I checked. (The test suite took nearly a day to run under Valgrind on a fast desktop machine when it worked, so is infeasible for every pull request, but could be done periodically, e.g. once every few days.)
Even a revived effort on getting core Julia tests to pass under Valgrind would not do much to help catch correctness bugs due to composing different packages in the ecosystem. For that, running in testing with `--check-bounds=yes` is probably a better solution, and much quicker to execute as well. (see e.g. https://github.com/JuliaArrays/OffsetArrays.jl/issues/282)
Comment · 5 years ago
> I'm giving Basecamp the benefit of the doubt
It's not clear that they deserve the benefit of the doubt, if we are to believe Casey Newton's article, "What really happened at BaseCamp":
> Employees say the founders’ memos unfairly depicted their workplace as being riven by partisan politics, when in fact the main source of the discussion had always been Basecamp itself.
> “At least in my experience, it has always been centered on what is happening at Basecamp,” said one employee — who, like most of those I spoke with today, requested anonymity so as to freely discuss internal deliberations. “What is being done at Basecamp? What is being said at Basecamp? And how it is affecting individuals? It has never been big political discussions, like ‘the postal service should be disbanded,’ or ‘I don’t like Amy Klobuchar.’
> Interviews with a half-dozen Basecamp employees over the past day paint a portrait of a company where workers sought to advance Basecamp’s commitment to diversity, equity, and inclusion by having sensitive discussions about the company’s own failures. After months of fraught conversations, Fried and his co-founder, David Heinemeier Hansson moved to shut those conversations down.
https://www.platformer.news/p/-what-really-happened-at-basec...
EDIT: formatting
Comment · 5 years ago
I have had great luck with "reMarkable Connection Utility." It bypasses reMarkable's cloud, which I have never once connected to. http://www.davisr.me/projects/rcu/
Comment · 6 years ago
I tried deploying a static site. DO App Platform makes the file page1/index.html available both at /page1 and at /page1/. If the page contains any relative links or images, it is impossible for both URLs to reference these links correctly. Most web servers handle this by redirecting /page1 to /page1/ before serving it. Netlify does this as well. Vercel works the same as DO in its default mode, but it allows the user to configure the trailing slash behavior to get the redirect. Hugo, for one, assumes all files are served with a trailing slash.
Comment · 6 years ago
It sounds like MDN could benefit from being the work of a 501(c)(6) business league, which is how the Linux Foundation's work is funded. Money could initially come from the big existing browser vendors. It is in everyone's interest to see it continue.
Comment · 6 years ago
Brad Smith in 2007: Linux violates 235 MS patents, and the FOSS community must pay up. https://web.archive.org/web/20070514143032/https://money.cnn...
Brad Smith in 2020: We Need a Hippocratic Oath for Software Engineers.
Comment · 6 years ago
Several years ago, I gave a presentation on quantum computing to the Los Angeles Hacker News Meetup. The slides are at https://jimgarrison.org/quantumcomputingexplained/ . Unfortunately, there is no video recording so they are currently lacking explanations.
My goal was to explain quantum computing in a way that is mathematically precise but doesn't require one to learn linear algebra first. To do this, I implemented a quantum computer simulator in Javascript that runs in the web browser. Conceptually (in mathematical language), in each simulation I present, I've started by enumerating the computational basis of the Hilbert space (all possible states the qubits could be in) and represented the computational state by putting an arrow beside each of them, which really is a complex number. (This similar to how Feynman explains things in his book QED.) The magnitude of the complex number is the length of the arrow, and its phase is the direction it points (encoded redundantly by its color). I've filled out the amplitude symbol with a square so that at any given point, its probability of a measurement resulting in that outcome is proportional to the area of that square. Essentially, in this language, making a measurement makes the experimenter color blind -- only the relative areas of the amplitudes matter and there is no way to learn directly phase information without doing a different experiment.
I could make a further document explaining along these lines if people are interested. The source is on github too: https://github.com/garrison/jsqis
I have prior experience working as a theoretical physicist in government and academia, and also as a software engineer. Before that, I worked as public relations coordinator at the Software Freedom Law Center.
On the web: https://jimgarrison.org/