@deadpantroglodytes's banner p

deadpantroglodytes


				

				

				
0 followers   follows 12 users  
joined 2022 September 05 13:29:17 UTC

				

User ID: 568

deadpantroglodytes


				
				
				

				
0 followers   follows 12 users   joined 2022 September 05 13:29:17 UTC

					

No bio...


					

User ID: 568

Google Search Keeps Getting Worse

I am hardly the first person to complain about Google search results circa Fall 2022, and I'm not the first person to recall how search used to feel like magic (1). It's become a commonplace (if a bit overstated) that for Google search to have any value at all, you need to point it at reddit.

Here's a case study: I've recently begun performing with a band so I went to look for ideas about how to improve my stage presence. Ten years ago, I would have just typed two or three words, perhaps just "guitarist live" or "watch guitarist live" and as I recall it, Google was reliably excellent at providing results that matched my intention, by either sorcery or science. Nowadays, perhaps superstitiously, I use complete sentences, so I typed "guitarists that are fun to watch live." The results were very bad. In order:

A group of video recommendations, all four suggestions useless:

  1. Rock Guitarist Live Streams For The 97th Time ! - With Guitar Solos, Chat, Games and Fun

  2. Three Chord Dave Live 50 Guitars music and good times

  3. Three Chord Dave Live 52 Guitars music and good times

  4. Three Chord Dave Live 51 Guitars music and good times

Next, a few links to articles:

  1. 13 Scorching Guitarists on Tour Today - Ticketmaster Blog

  2. The best live streams and virtual concerts to watch while social distancing (take note of this one, from April 23, 2020: we'll come back to it later)

  3. and 8. The next results were the "People also search for" and "People also ask" suggestions, none of which were helpful.

  4. A link to an Insider article called "Musicians you need to see live in concert". That sounded promising, except none of the musicians were guitarists (none of the headline musicians advertised on the search results - once you drilled in, Lenny Kravitz and the Red Hot Chili Peppers probably qualify).

  5. A youtube video, "Top 10 Guitarists of All Time (REDUX)". Closer, but not really what I'm looking for.

More useless results followed, including three of the next five focused on streaming ("A Guitarists Guide to Live Streaming").

Finally, coming to the culture war angle, I want to ask why this might be. Why are Google search's results so bad, compared to the five or ten years ago, or even farther back, when they had inferior technology? Clearly, some of the problem is spam, as many people argue. But that doesn't really decribe what I saw. Is it because they are prioritizing social justice in results? I know this flatters the Motte, but it also explains a handful of the noise in my search above, on a fairly anodyne topic. I got three results about streaming performances: sure, maybe "live" is often linked for "livestream" (Plato's pharmakon strikes again, three cheers for auto-antonyms!), but what explains the second non-video recommendation, number 6 above, "The best live streams and virtual concerts to watch while social distancing"? In my mind, I'm trying to find tips about how to perform live on stage for people in sweaty clubs, gleefully exchanging airborne microbes, and Google's trying to shove an article from April 2020 down my throat. I couldn't do better if I tried to parody this.

If you think my expectations are crazy, I get it, except until recently (geologically speaking), Google would have delivered EXCELLENT results on this topic.

Some other possible explanations:

  • SEO has gotten better than search - this could explain some of what I saw.

  • The internet is crowded now, there's more surface to search! That doesn't seem likely - certainly not substantially more so than five or ten years ago.

  • Maybe google never was magic! I have a bad memory or it just seemed incredible because it was novel.

  • Goodheart's law / overfitting, definitely part of the story: optimizing for revenue reduces engagement and relevance. But then again, so does optimizing for justice! It's hard not to suspect how the often comical and heavy handed attempts at "alignment" have marred ChatGPT.

  • Google engineers are bad. Non-starter, based on the ones I know. Google has lost a lot of excellent people over the years (like Steve Yegge, etc.) but this doesn't add up.

  • Google hires good people, but they don't funnel their best talent into search, because they continue to have an effective monopoly, even in the age of Bing, Duckduckgo, and Kagi.

  • ???

Which is it?

1 https://freakonomics.com/podcast/is-google-getting-worse/

The quoted comment says the Mongols' genes were successful in the appears to say that their genes prevailed, not that their genes caused their success.

Not to come down too hard on what might be an off-hand comment, but exactly what salient or urgent "danger" does fake history present?

I can see how historical narratives (fake or not) could motivate conflict. For example, stories of dispossession can establish a deeply felt grievance that can be pointed to various destructive ends. But:

  • Not all dangerous narratives are in fact fake.

  • Those stories can just as easily be pointed at various constructive goals.

  • Most history isn't emotionally inflammatory.

  • Our understanding of history is constantly undergoing revision. Do discarded historical narratives count as fake? Are they prima facia "dangerous"?

  • The stories are open for (and usually attract) rebuttals.

When I was a kid, everyone read von Daniken's Chariot of the Gods, spent a few weeks mind-blown, then moved on to adulthoood.

Isn't the real danger here, the great, recurring theme of the internet era, the crisis of authority?

A well-deserved crisis, I'd say, but also genuinely dangerous to a great many people.

I enjoyed a lot of this, and admired Shakesneer's composure under rigorous questioning even if I think you have the better of the argument, Yassine.

But wow, I wish I'd skipped the first hour. Couldn't you just stipulate that the PMC hates MAGA voters instead of pressing so hard on their motives, in this context? For a while, I thought you might be trying to go Socratic and lure Shakesneer into admitting that the Feds have a rational reason for persecuting conservative groups, beyond losing their jobs (if one grants that they persecute them). But it didn't quite cash out that way.

I'd say that employees of federal agencies have strong motives for their hatred of conservatives, even setting aside the "I can tolerate anything but the outgroup" reasons and fear of losing their jobs, especially if we're talking about the FBI and ATF:

  • Conservative political thought emphasizes the contingency of the state's legitimacy, moreso than the left (CHAZ notwithstanding). Right-learning separatist groups are closer to the mainstream of the conservative movement than those on the left, at least in the US.
  • Conservative political thought challenges the state's monopoly on violence. Pro-gun advocacy makes the FBI and ATF's jobs harder.

And that's how it plays out in real life. I have several friends that are FBI agents, and I occasionally go to parties where more are present. To a person, they take it for granted that Trump and his voters are contemptible. I'm sure it's not unanimous, but it definitely isn't generational - I'm a Gen-Xer, and the parties are generally +/-10 years around me.

Turchin seems to believe in, for lack of a better term, socio-economic Malthusianism, and his more formal historical work requires Rube-Goldberg-style epicycles to substantiate his grand theory (constantly zooming in and out of geographic regions and gerrymandering timescales to make data fit). But even even though his Spengleresque ambitions won't amount to anything, "elite overproduction" is an exceptional framework for explaining local conditions in varied social environments, like the American media or academia. I can't see any reason to take him more seriously than that.

This was the justification for affirmative action 1.0, and is occasionally still evident as a first line of defense, but aa 2.0 is based on two completely different ideas:

  • That diversity makes organizations stronger in a variety of ways.

  • Proportional representation is required for organizations to be "democratic", in the somewhat novel sense of engaging the whole population.

I agree that this is an old phenomenon with a long history: courageous teachers becoming involved with a child's welfare at some risk to themselves. But institutionalizing it changes everything. Guaranteeing state support dramatically reduces the risk to the teacher, which destroys the balance of incentives.

I'm sympathetic to kids trapped in a hellish adversarial relationship with their own parents, but predict that solving their problems by substituting state-approved parental figures will create a different series of problems that will probably affect a much larger number of children. Attempting to solve a tiny minority of problem cases, these laws create a new vector for neglect and abuse, because they cut parents out of the loop, when they are, in most cases, the people most committed to a child's well-being by many orders of magnitude.

Same world, different screens? I don't know how to reconcile these two comments with my personal experience.

My spouse has been a tenured humanities faculty member at a BAU for 20 years, with several different stints across the country the decade before that. Between our own experience and that of dozens of friends in the academy, everything the OP wrote rang true to me, except the timeline at the conclusion (our institution is 5-10 years ahead of the OP's account).

I'd add that faculty social life is stultifying - it's not that you can't ever have real conversations with people about difficult topics, but it takes a long time to break through the suffocating blanket of conformity. Most social encounters start with progressive consensus-building about the issues of the day, and often can't move past that. It's worse if there are unfamiliar people in the group, or administrators.

These phenomena may not universal, but are, at a minimum, widespread. Above all, I'd love to know what your institutions are doing right that you don't see this.

It has absolutely lost its original propagandistic edge and become all-purpose filler. A recent NYT article about a scandalous Russian party made me laugh out loud:

The suggestive photos and videos that surfaced on social media soon after were unremarkable. Yet the blowback was immediate and severe.

“The country is at war, and these scum, beasts, are putting on this,” one of the country’s most prominent propagandists, Vladimir Solovyov, wrote on his Telegram channel hours after the event. “Cattle who don’t give a damn about what’s happening.”

Some prominent conservatives went further, claiming, without offering evidence, that the party was a satanic ritual because it occurred, according to their calculations, on the 666th day of the war in Ukraine.

Perhaps it was a joke? Is anyone here willing to own up to entryism at the New York Times?

[Edit - instead of being so snarky, I feel like I ought to give more weight to the opposite interpretation, that it was an occult signal, a quiet protest or a cry for help from within.]

The Democratic response is infuriating. The way this is currently playing out is simple: people and families that would benefit from skilled nursing either get no care or, like a close relative of mine, spend weeks in a hospital (soon, months).

As an aside, this has increased my appreciation of the Fed and reduced my enthusiasm for keeping unemployment extremely low as a method of spreading prosperity.

Hanania should just say that his views have changed, and that he is no longer the kind of angry young man who wrote those comments

Bullseye.

I've been with you all the way up to here, upvoting and even reporting one of your comments in this thread as a QC, but this is clearly wrong: the feds can be guilty of luring the protesters/rioters into taking risks that would make them look bad and expose them to prosecution.

I was vehemently agreeing with this comment, until the unfortunate conclusion:

I trust that combination of judgement [the student's and school's] a fairly high amount, again especially given the asymmetric dangers involved here

This idealized view of educational personnel doesn't stand up to scrutiny. It's mired in the present conflict and lacks perspective: institutions have a long history of abusing the trust of children and do not deserve this level of confidence or deference.

It's tawdry to quote myself, but I don't think I can put it better than I did before:

I agree that this is an old phenomenon with a long history: courageous teachers becoming involved with a child's welfare at some risk to themselves. But institutionalizing it changes everything. Guaranteeing state support dramatically reduces the risk to the teacher, which destroys the balance of incentives.

I'm sympathetic to kids trapped in a hellish adversarial relationship with their own parents, but predict that solving their problems by substituting state-approved parental figures will create a different series of problems that will probably affect a much larger number of children. Attempting to solve a tiny minority of problem cases, these laws create a new vector for neglect and abuse, because they cut parents out of the loop, when they are, in most cases, the people most committed to a child's well-being by many orders of magnitude.

I had a lot of great teachers, people that encouraged and supported me, but I also had egomaniacs and narcissists who took great pleasure in driving a wedge between kids and their parents (with no long-term concern for the children). I saw more than a handful of teachers happy to sexually exploit their students*. And I saw a substantial minority of lazy, time-serving clock-punchers.

* And a few relationships that I wouldn't call exploitative, but imagine most would.

I agree, not much, but the federal government can definitely avoid exacerbating the problem by limiting supply.

It seems quite easy to interpose a cut-out for crypto-to-fiat conversion, so that your only real exposure is to tax authorities. The only challenge is ensure there are enough socially acceptable crypto use cases that you don't draw too much attention.

I wasn't bored, but as a movie it sucked: it's an artless piece of relentless exposition, a paint-by-numbers Wikipedia adaptation. Nolan doesn't understand the most important unit of cinematic language, the scene, and tries to hide it with garish sound design and 60s-era hallucinations.

It had at least three times too many characters as a movie can really handle. He should have started by cutting Oppenheimer's brother, who adds nothing substantial to the film. If there's no way to tell the story without them, Nolan should have just told a different story, one better suited to the strengths of motion pictures.

I also prefer the written word, but I like to listen to podcasts when running and driving.

Podcasts also give me an option to avoid the farce of trying to do dishes while thumbing through essays and comments with wet fingers.

Plain old supply and demand drives CEO pay, for the most part, not value-over-replacement.

Even sub-replacement-level CEOs require an extremely unusual skill set. Large companies are like Game of Fucking Thrones. Running one is like herding cats where one of every three animals is actually a serial killer, a thief, or a company-destroying liability risk and three out of four would sell you for a sandwich, must less a shot at your job. On top of that, you're steering (at best) half-blind in a hurricane, an earthquake, and a tornado at the same time (macro-economically speaking).

There are certainly problems, like back-scratching boards determining compensation, but I don't see obviously superior alternatives.

Why should there be such a middle ground? "Cancellation" aims to make someone unemployable, which is several steps short of murder but absolutely moving in the same direction, and not in the slippery-slope sense.

Yeah, I don't think there's a plausible story to explain a widespread deep-state conspiracy. But there are other plausible theories (completely unsubstantiated, to be clear). They would involve either

  1. an informant with bad judgement going full Leeroy Jenkins or
  2. an undercover agent assessing the crowd, deciding they didn't pose a real threat, and doing their best to goose the crowd on

In the absence of evidence, those are idle speculation, but I'm a little surprised they didn't come up (or at least not more clearly).

As another data point, in my affluent blue bubble, hyphenated children's names are everywhere, partly because of the relatively large number of same-sex couples, but not remotely exclusively.

This includes mine, and while I have to admit that hyphenated names can be a mouthful (paving the way for endless jokes about fitting names on, e.g., swim caps), I love them for surfacing family history. I've always loved the extravagance of Spanish names for that reason, even if they are usually truncated for daily use.

As an aside, I appreciate the link. I've been listening to Huberman a bit lately because I'm starved for good advice about how to deal with an alcoholic family member, and it's been very hard to come by a trustworthy assessment of his podcasts (like Ritchie).

I wrote "we should not seek the destruction of people that disagree with us (neither literal, social, nor professional destruction)". Your response:

Is the walk of shame supposed to be social destruction?

Were you intentionally dodging the substance here? I don't even have to speculate on their intentions with respect to social destruction: the law students are calling for their teacher (the dean of the law school) to be fired. If you don't agree they're trying to destroy her professionally and that the walk of shame was part of that, we don't have enough common ground to discuss this.

Presuming that was an oversight, I agree there has to be a gradient of offenses and responses. There's an entire universe of proportionately calibrated responses that don't involve silencing or attacking the speaker:

  • Ignore them.

  • Participate in the Q&A, ask sharp questions.

  • Organize a local event featuring a speaker or speakers providing a counterpoint.

  • Publish something critical of the ideas.

I'm aware that people often characterize boycotts, de-platforming, and collective shaming as an alternative to violence, but I think the opposite is true: these things all escalate towards violence. Their widespread currency fuels the volatile, scary environment in which we live. I would prefer to see our society establish different norms that would support engagement and follow the examples of Ira Glasser and Daryl Davis.

Edit: "walk of shame wasn't part of that" -> "walk of shame was part of that"

I oppose the "walk of shame" on the grounds that it is intimidation.

My principle is that we should not seek the destruction of people that disagree with us (neither literal, social, nor professional destruction) unless we're willing to kill or die on the relevant hill. (In case this comes off as melodramatic, I don't mean this in the "I'm an internet tough guy" sense, but nearly the opposite: I think few hills are relevant in this way.)

Edit: "nor", not "or" and added parenthetical clarification.

Here's a longshot: I'm trying to see if anyone else read (or wrote!) a comment on the National Review website ... from a few years ago.

A few months after Trump won the election, I started reading NR's investigation coverage. I stopped when they started paywalling Kevin Williamson, so the comment would have appeared somewhere between April 2017 and July 2021. The comment was an extremely persuasive compendium of offenses the mainstream media had committed against red America. In retrospect, I realize that it might have been copy/pasted from the motte, or perhaps written afresh by someone here (I didn't start reading the motte until 2020).

Does this sound familiar to anyone, or does anyone know of a similar resource?