mttpgn

Clarification: when Aaronson quotes Gur to say that Kristof is "feigning that he can’t see the water," that line is a callback to an earlier (unquoted by Aaronson) sentence in Gur's original post: "Fish don't notice the water."

Thanks, and I got the main window open now, but I'm getting a second error that doesn't look OS related. `Plugin error (txtfilter.py): No module named 'miniword.importexport'`

On MacOS, I'm seeing `ModuleNotFoundError: No module named 'miniword.core.utils'` whether I run `python3 -m miniword` from src/miniword/ or from src/miniword/miniword/.

There are many computing platforms that rust devs simply cannot be bothered to support (for just one example, Cygwin). The Python devs, by contrast, have actually put in a lot of work over the years to support a plethora of operating systems and architectures.

Requiring Rust as a CPython dependency would be to abandon all their users stuck in obscure environments.

Your site has a search bar for typing in a full prompt to an LLM about what is my current mood, and I just find it interesting that one's mood is the important thing for your users to supply as input to your service. For me, unless a major event has taken place, I usually don't take time to think much about what's my mood beyond one or two words. If I've been on a journaling kick I'll usually write about the concrete experiences of the day as a proxy for describing my mood without actually getting to what this means for my energy levels/affectations, etc. The mood descriptors I do recognize in myself (eg. kinda sad!) generally factor little into my content consumption decisions (at least consciously). More important to me are questions like "What are folks talking about? (driving discourse online or at the office)", "Which movies have been recommended to me (by friends/family or by advertising)", and "What's accessible? (On a service I already subscribe to without needing an additional purchase)".

Pandas code is untestable

The thousand-plus data integrity tests I've written in pandas tell a different story...

The BBC published this article. I agree with "all of literature" being hyperbolic though.

Phillip

I too have a `mkcd` in my .zshrc, but I implemented it slightly differently:

  function mkcd {
    newdir=$1
    mkdir -p $newdir
    cd $newdir
  }

So I prompted my brain

Is there anyone for whom this phrasing has a clarity advantage over "So I asked myself..."/"So I thought to myself..." ?

Ditto

This submission links to the actual paper. The other submission is the NY Times' summary of it.

What a great reply this graph can make the next time we see VCs on X complaining how, "Everyone I talk to has so, so many engineering positions they just can't fill". By these measurements, the job market looks worse now than it did at the troughs of lockdown.

An excerpt from _How to Lie with Statistics_ by Darrell Huff (1954):

Take this one: "27 percent of a large sample of eminent physicians smoke Throaties--more than any other brand." The figure itself may be phony, of course, in any of several ways, but that really doesn't make any difference. The only answer to a figure so irrelevant is "So what?" With all proper respect toward the medical profession, do doctors know any more about tobacco brands than you do? Do they have any inside information that permits them to choose the least harmful among cigarettes? Of course they don't, and your doctor would be the first to say so. Yet that "27 percent" somehow manages to sound as if it meant something.

That book specifies many other examples (from this time period in America) of misleading claims that sound statistically significant upon an uncritical, cursory reading.

To be fair to Zweig, history's events could have unfolded differently than they did.

Joe-Nathan? Yonatan?

The position doesn't have to be "in the money" to be profitable. At present, the underlying (that is $CRWD, not the puts themselves) is trading around $305 per share, so these puts with the strike price all the way down to $185 are still far, far off from being ITM. But, options are largely a bet on volatility. While the extreme price drop between when this user opened the trade and this morning has already made this trade quite profitable, it could still lose money if the trader holds on to it; IANAFA, but selling and taking profits on this rare event would most likely be in their interests, even if they continue to believe the underlying security will go all the way down to $185 by November.

I want to agree with the point you're making, but WannaCry, to take one example, had an impact at roughly this scale.

Yes, elevenlabs can.

I find myself grateful I can support my family financially, as it's statistically unlikely my art can ever permit me to quit my day job.

Not to imply that it's in any way a quality peculiar to those who write poetry in their free time (I know it isn't), but my need for creative expression regularly seeps into the coding standards to which I hold myself, insofar as I demand a certain beauty-- such as such a thing is ever even possible-- out of the libraries I write.

I build and deploy business process automations to AWS using Python.

The link directs to an announcement via "toot" on mastodon.archive.org

To modify tags on MP3 files programmatically, I have found the Mutagen library works well in Python.

Licensing the emotion-intoned TTS as a standalone API is something I would look forward to seeing. Not sure how feasible that would be if, as a sibling comment suggested, it bypasses the text-rendering step altogether.

I have never found protonmail's signup step asking for phone number verification or a recovery email to be unskippable.

Protonmail can still be the best choice for a pseudonymous mail service so long as it's combined with diligent, consistent IP address obfuscation. Protonmail will continue to allow logins and new account creations over Tor. All the major free email providers have long since disallowed new signups over Tor, and most have some form of degraded user experience when logging in over Tor, if they allow it at all. Small, niche email providers appear and disappear so often that relying on them still to exist even a few months into the future is a big gamble. Hosting one's own email requires payment of some type to the hosting provider, so it is not anonymous. Other privacy-oriented free email providers, such as riseup, will do exactly what protonmail did, because if they refuse, their only option is to go the way of lavabit.

The cryptographer Dan J. Bernstein once told me a story that Bruce Schneier kept some cryptographic protocol secure for an additional 24 hours. The researcher demonstrating this protocol's weakness based their proof-of-concept on a proof in Schneier's book. However, Schneier's description contained a mathematical error. When the error in the proof-of-concept was pointed out to the researcher at the conference, this researcher went back to their hotel room, discovered the origin of the error in Schneier's text, and fixed the proof-of-concept for the conference-goers by the following day. Thus, Bruce Schneier kept a cryptographic protocol secure for an additional 24 hours.

For an overview of other complexity classes besides P and NP, there's https://complexityzoo.net/

The adoption of this motto by a company employing thousands of people provides a great example of why we need more engineers versed in the humanities.

What is evil? If it's violations of the law, then why not "Don't break the law"? (That certainly gives off stronger "already answered my t-shirt" vibes).

Does 'Don't be Evil' mean not violating any principles of Sartrean humanism? Those are all perfectly clear, right? Does it mean keeping the Judeo-Christian decalogue? The Confucian analects? Good thing there are no ongoing debates problematizing those! Or maybe it requires performing Utilitarian harm reduction calculations before all decisions? That's easily calculable, though, isn't it? Or, perhaps it involves abstaining from all meat, alcohol, and cigarettes? Some Jains wear special shoes to avoid accidentally crushing insects while treading, while Mormon ethics require the donning of special undergarments. Will corporate provide either of those?

Aside from the fact that the existence or non-existence of metaphysical evil is still debated both in secular-rational and theological contexts, it's unavoidable that many ethical ideologies fundamentally contradict one another. One can either choose one and rule out all others, hoping that the chosen model proves internally consistent and unambiguous, or one can accept that, as Philip K. Dick once said, "We are all doomed to commit acts of cruelty or violence or evil; that is our destiny, due to ancient factors. Our karma." Neither choice sounds ideal!

Friedrich Nietzsche's theory on the inevitable "transvaluation of all values" lends some weight to the impossibility of retaining this maxim in Google's corporate code of conduct forever. Reading into Google's abandonment of 'Don't Be Evil' this uncharitable idea that leadership has decided "evil" (whatever that is) is ok now, actually, leaves one with the unsatisfactory feeling of missing some perspective.

Instead, I think, the decision shows us that even the most simply worded of slogans, with the best of intentions, are inadequate as frameworks to think deeply about the often impossible contradictions of our beliefs and commitments. Ethical action requires deep thought. It's not a consensus, and it's not majority rule either (probably!). Not doing evil is complicated, and hard. It's not something that every preschooler inherently understands at playtime. And even if we follow through on the most informed decisions, supported by the all-true suppositions and ironclad logic, we may still find ourselves doubting that course of action years down the line as new outcomes are disclosed.

In the late 1990s, my dad attended night classes with other adult learners to earn his MBA. Everywhere he went during those years, my dad had a beeper clipped to his belt for an on-call hospital rotation. During his first week of class, my dad's beeper loudly went off during the middle of the lecture. As my dad scrambled out of the lecture hall to call the number on the little screen, the professor accurately guessed: "You're a doctor, aren't you?"

Just because the number of abstraction layers can be reduced doesn't mean they need to be. You might gain back some CPU cycles, some milliseconds of execution time. But the tradeoffs of maintainability, legibility, and developer quality-of-life may, in the long run, reintroduce abstraction layers of some other type back into the overall SDLC.