Inkwell: Authors and Artists

    
I find plenty to read on BSky, very little to interact with. But I
really enjoy seeing the people there who used to be on Twitter. They
are not friends, but still people who presence I value.
  

    



David Coote writes:



In permalink #<170>

"Write 50,000 essays on a near daily basis over 25
years"

That would be 2000 essays/annum.

Assuming 200 workdays/year that's 10 essays/day!

Which sounds a tad high.


And I've been reading these SOTWs for years.

I'm finding this one quite boring.




<https://people.well.com/conf/inkwell.vue/topics/540/Bruce-Sterling-and-Jon-Leb
kowsky-page07.html#post170>





(note: conference host added link)
  

    
An Italian programmer who goes by “antirez” wrote about AI in
English:

Translating blog posts with GPT-4, or: on hope and fear
<http://antirez.com/news/141>

And in Italian, then using GPT-4 to translate it into English:

LLMs and Programming in the first days of 2024
<http://antirez.com/news/140>

> Another point of view on the matter could be that my true voice is
the one of the translated blog post, so writing in English is the
real bluff here. Because the translated post is more representative
of my lexical ability in my mother tongue, and not of the reduced
one I can feature when writing in English. Maybe it captures more
shades of what I really want to say. But then one could go deeper in
arguments about what style really is. Is it more about sentence
construction, and the way you put down your ideas, or is it a lot
more about the vocabulary used, the exact words and adjectives
selected to provide a given image and meaning? Probably both, and
the two things are quite an inseparable whole.

> Anyway the simple fact that now, in 2024, I finally have this
choice, fulls me of hope and fear. Hope for the possibilities the
humanity will have, with machines that can talk. And fear about the
potential AI has to make everybody lazy, no longer willing to do
things as hard as learning a new language.
  

    
(The quote is from the first linked post, written in English.)
  

    
And here is a quote from the second linked post, written in Italian
and translated:

> It's true: LLMs are capable, at most, of rudimentary reasoning,
often inaccurate, many times peppered with hallucinations about
non-existent facts. But they have a vast knowledge. In the field of
programming, as well as in other fields for which quality data are
available, LLMs are like stupid savants who know a lot of things. It
would be terrible to do pair programming with such a partner (for
me, pair programming is terrible even in the most general terms):
they would have nonsensical ideas and we would have to continuously
fight to impose our own. But if this erudite fool is at our disposal
and answers all the questions asked of them, things change. Current
LLMs will not take us beyond the paths of knowledge, but if we want
to tackle a topic we do not know well, they can often lift us from
our absolute ignorance to the point where we know enough to move
forward on our own.

> In the field of programming, perhaps their ability would have been
of very little interest up to twenty or thirty years ago. Back then
you had to know a couple of programming languages, the classic
algorithms, and those ten fundamental libraries. The rest you had to
add yourself, your own intelligence, expertise, design skills. If
you had these ingredients you were an expert programmer, able to do
more or less everything. Over time, we have witnessed an explosion
of frameworks, programming languages, libraries of all kinds. An
explosion of complexity often completely unnecessary and
unjustified, but the truth is that things are what they are. And in
such a context, an idiot who knows everything is a precious ally.

> Let me give you an example […]
  

    
As a programmer (retired) I’ve dabbled a bit and had a similar
experience myself. Here’s a Python notebook implementing a bouncing
ball animation, using Python API’s that I didn’t know at the time.

<https://github.com/skybrian/boing_ball_animation>

Ironically for a field called “machine learning,” a frustration when
using these tools is that the chatbot can’t learn from experience in
any permanent way. In the next chat, it’s going to make the same
dumb mistakes that it made last time. (Until the vendor releases a
new version.) A co-worker would learn from practice, but not the
machine.

When you know very little about a subject, you start out thinking
it’s quite helpful, but as you learn more, it becomes less useful,
and then a hinderance. But there are always more things to learn.
  

    
I happened to return to the Spanish kitchen-supply store that I
mentioned earlier  (post #114), where they sell that dieselpunk
glassware which mimics the Roaring 1920s but is actually sold here
in the enshittified  2020s.

With a closer look, I discovered that this crystalware is actually
"Made in Italy,"and then it's shipped to Spain so that both the
Spanish and Italians can mock the Prohibition ambitions of the
Americans.

And it's better than that: the "new America '20s collection" is
described on the carton in seven different European languages, so
that practically everyone in Europe can laugh at the very concept of
legislated sobriety.

That discovery tipped me over.  I just bought half a dozen of those
glasses.
  

    
Pics or it didn't happen.  This is not an Erskine Gwynne
"Bopulevardier," but it is an Ernest Hemingway "Hemingway Daiquiri,"
and since Hemingway used to write for Erskine Gwynne, it's very
period-appropriate.

https://flic.kr/p/2pryfrc

It's also made with some Hemingway-centric Cuban rum, which is
sternly prohibited in the USA but as common as lime juice here in
Ibiza.
  

    
Back to the quality of online experience. I think an interesting
question in this context is: why are Google and other major search
engines now curtailing their search results?
  

    
Can you day more? What do you mean by "curtailing"?
  

    
Fewer pages of results among other things. Less reach into past
articles and references. 
  

    
This might be at least a partial explanation:
<https://searchengineland.com/google-expect-more-simplification-of-search-resul
ts-432305>

"One of the three key points of emphasis in [Gary] Illyes' keynote
was that Google has been removing SERP features that Google users
didn't use or find useful. That include the change to video
thumbnails in April and then the recent removal of HowTo rich
results and heavily reduced visibility of FAQ rich results."

This piece discusses how Google's search results have been changing,
including an AI result at the top of the SERP:
<https://searchengineland.com/new-google-search-generative-ai-experience-413533
>
  

    
An explanation maybe. But the quality and utility have clearly
declined. 

There’s also been increasing manipulation of which results show. For
example, the left-wing publications Common Dreams, AlterNet, and
Counterpunch have all had serious problems with either search or
social media access.
  

    
Can you link research with more detail about these phenomena? I'm
not sure how to interpret what might be happening without more
detail about what the studies are actually reporting. Search
algorithms are always changing, sometimes with devastating results
for some entities. I recall when medical sites experience
substantial loss of traffic after a change to the Google algorithm a
few years ago.

I recently started using DuckDuckGo, and find results there pretty
clean and straightforward.
  

    
Interesting. I should be clear that the two things I mentioned in
<188> are two separate issues.
  

    
Seems like there’s more interest in alternative search engines than
there has been in a while. It’s unclear if any are better. Fans make
claims. There are reviews, but the ones I’ve seen seem pretty
lacking in methodology. And anyway, everyone searches for different
things, so you have to evaluate it for yourself, and that’s a drag.
Easier to stick with habit.

AI chatbots have the same problem. There are dozens of these things,
more being announced all the time, and it would take genuine
curiosity and serious commitment to try them all. And how do you
evaluate them? There is no good IQ test for bots. Sure, you can test
them, but they are idiot savants whose flaws aren’t easily
summarized in a single number, or even a few dimensions. There are
lots of papers and lots of benchmarks, but the benchmarks results
don’t seem so meaningful for everything tools like search engines
and chatbots. (Machine learning researchers have fun trying to beat
each other on benchmarks, though!)

Sometimes I wonder if Google has lost its way because the evaluation
problem has finally done them in. They have an army of human
evaluators, but they’re just ordinary people who can be fooled. Do
they struggle to tell good search results from bad? You can’t tell
by a glance anymore, particularly in areas outside your expertise.

I’m relying on social proof, hoping that other people will try
things and there will be buzz about anything particularly promising.
Until then, I use GPT-4 and Google. It’s what I’m used to.

Relying on buzz is pretty terrible too. Take human evaluation by
biased partisans who probably believe a few conspiracy theories, and
then run through social media algorithms. But it’s the lazy way out
and that’s appealing.
  

    
Ah also interesting. I should also clarify that I could care less
about the mechanisms involved. Search is just another black box. I
only care about the quality and consistency of my personal user
experience as I’m sure do many others. If the algorithms are
constantly being tweaked that affects the latter. Search should have
some sort of baseline consistency otherwise it’s functional
reliability is compromised. I'm stating the obvious but
nonetheless....
  

    
I just read that Vienna has the most affordable housing because the
city itself owns and manages a big percentage of the housing stock. 
Apartments rents there are much lower. But the city is slow to
respond to maintenance problems.  

https://www.theguardian.com/lifeandstyle/2024/jan/10/the-social-housing-secret
-how-vienna-became-the-worlds-most-livable-city

Socialism at work.  Too bad it's such a dirty word over here.
  

    <scribbled by jonl Wed 10 Jan 24 09:47>
  

    
Via email from David Coote:

In permalink <170>

"Write 50,000 essays on a near daily basis over 25 years"

That would be 2000 essays/annum.

Assuming 200 workdays/year that's 10 essays/day!

Which sounds a tad high.


And I've been reading these SOTWs for years.

I'm finding this one quite boring.
  

    
If anybody can write 10 essays per day, it's <doctorow>! 

As for whether the conversation's been boring - I don't agree, but
that's just me. I wonder what other readers think. 

My hope was that we'd be insightful, not necessarily exciting.
  

    
Re: #172

Thanks for sharing your knowledge and insights, Isaac.
  

    
Seconding those thanks.
  

    



Rick Crain writes:

David Coote's skepticism of Cory Doctorow having written 50,000 essays in
25 years would be lessened had he read Doctorow's work during his many
years as a BoingBoing contributor. Now it seems that he just writes one
really in-depth essay for pluralistic.net most days, plus still a fair
amount of others too.

Rick Crain
  

    
wrt #195, maybe this sotw is boring because the actual SOTW is
indeed boring. bad and boring, for sure.