My action at Academia.edu over the last 6 months

Has been pretty interesting. Look at these charts, for January 2026 through June, in order from top to bottom:

January is pretty even, with a peak just below 50 near the end of the month. February shows a steady rise through the month, topping out just over 60 at the end of the month. March tops out just over 70 at the beginning of the month then takes a wild ride though the middle finally evening out a bit at the end of the month, with two peaks just above 60 and on just below. April has its ups and downs but on the whole seems pretty steady, with values between just above 20 to just above 60; 40 seems to be the middle value. May shows a definite downward trend, with a high of about 55 about a week in and then dropping through the end. And then we have June, which starts out low rises to 60 in four days, hangs in there a bit and then starts dropping down until the last day (June 30) it has a steep rise until well above 90. Then an apparent steep drop to July 1, though that’s an illusion. I took that last screen shot very early in the day.

I have no idea what that steep rise is about. My guess is someone somewhere put out notice about me – email, social media, website, who knows? – others picked up on it and came by. The question is: what’s next. It’s now early in the morning on July 4 (7:30) and here’s what we’ve got:

We’ve dropped way down, down to the region where things were for the middle through June 29. Note though, that we’re well above 15, which is pretty high for this time of day. Still, I don’t expect a steep rise through July. Beyond that, who knows?

Interesting action at Academia.edu

Here’s four successive months:

January 2026:

It’s relatively constant, with a peak at 49 and three troughs at 10.

February 2026:

There’s a rising trend through the month with the peak above 60 hits at the end of the month and one just below that four days earlier.

Now March 2026:

It’s pretty much up and down, but there’s an early peak above 70 and things seem fairly steady for the last week and a half. A low dips down to 10.

April 2026:

It seems fairly high on the whole. Three peaks above 60 distributed over the month, with a trough at 21 and a later one at 22-23.

Throughout this period I’ve been in the 99.5 percentile. How long before I break 99.9?

By the numbers, New Savanna and Academia, a quick update

Here’s where things currently stand with overall views at New Savanna:

There’s that big jump for 2017. I haven’t got the foggiest idea what that’s about. Nor do I understand the big rise that’s been going on since late 2023, but at this point it’s been going on longer than the previous one. I have no idea what’ll unfold over the next year.

Now look at this, which compares my overall performance on Academia.edu with Steven Pinker’s:

We’re both in the top 0.5% of views for the last 30 days. The last 30 days. That would seem to imply that I’m running neck in neck with Pinker, which is absurd. He’s one of the best-known intellectuals in the world, while I’m unknown. He’s got 75 times the number of followers I have and 2.9 times as many total public views. But for the last 30 days, there we are.

Report Card: New Savanna and Academia.edu, end of 2024

Time to report on traffic here at the Savanna and over at Academia.edu. As of Jan. 1, 2025, this is the traffic at New Savanna over the past year.

Now look at all-time traffic:

Notice that big rise in 2017. I haven’t got the foggiest idea what happened there, but whatever it was, it didn’t last. Traffic was back down, though perhaps just a big higher than before that period. Then we got a big jump at the end of 2023 and that stayed with us through 2024. I don’t know what that’s about either, but it does seem to be dropping off. How far down it’ll go, I don’t know.

Here's the top posts over the last year:

The top two have been riding high for a long time. The second one, about Kawajiri’s Ninja Scroll, is one of the oldest posts here, and it’s about sex, a surefire winner. The third post, Shakespeare on the wane? went up in October of last year. The reason it has so many hits is that Tyler Cowen mentioned it over at Marginal Revolution, which has a huge following. Similarly, GOAT Literary Critics went up on December 4, 2023. It also got a boost from Cowen. But the last one down, Claude 3.5 Sonnet visits Times Square, didn’t get a boost from Cowen. It sent up on December 9, 2024 and got over 600 hits on its own (the tallies in Blogger’s multiple-post listings tend to be lower than the tallies you get at the posts taken individually). I’m not sure why this one got so many hits, way more than any of the other “Claude describes” posts.

Now, to Academia.edu. At the beginning of the year I was, I believe, at the 99.5th percentile in total hits, however it is that Academia measures that (I believe it’s over the previous 30 days). I’d been as high as 99.9 sometime during 2023. It dropped to 99.0 during the middle of the year, but now I’m back up to 99.5.

Here's the traffic for the last 60 days. The green line is paper views and the black line is unique visitors. The blue line is for profile views.

This table is sorted by most all-time downloads in descending order (rightmost column). Notice the document with the most downloads is a paper with practical advice on how to play relatively simple polyrhythms on a tongue drum. An encyclopedia article on Visual Thinking is third.

Keeping Score of my online action

It’s a lazy Friday afternoon, but then recently all Friday afternoons have seemed lazy. I’ve spent much the day doing this and that and taking stock, getting ready for the been weekend push on my next piece for 3 Quarks Daily. It’s currently titled, “Why We Need Philosophy, Now, More Than Ever [AI].” Perhaps it will have that title by the time I publish it. We’ll see.

Anyhow, I thought I’d report my latest states on Academic and here at the Savanna.

Academia

I’ve got 90,841 total views, which puts me in the top 0.5%. I’ve been there for months. At one point in the last year, Mar. 22, I was in the top 0.1%. That’s pretty good. I need to be in the top 0.001%.

The chart below shows my paper action over the last 60 days. The upper green line is paper views and the lower black line is downloads. Notice that all the data points except the rightmost one indicate totals for the day. I took the screenshot at 4PM; I won’t hazard a guess where thing will be at the end of the day.

Here's the most popular papers over the last 60 days (click on chart to embiggen):

Notice that the top paper, on the tech Singularity, is way behind the second paper, on Kawajiri’s Ninja Scroll, in total views (to the right), but ahead of it in downloads. The singularity paper is from 2014; the Ninja Scroll paper is from 2015. The corresponding Ninja Scroll post at New Savanna is the most popular post there. I’m particularly pleased with the performance of the Heart of Darkness paper, which is hard-core literary criticism, with quite a bit of close-reading and “distant” reading as well; it’s from 2019. The Visual Thinking paper is an encyclopedia article I published in 1990.

New Savanna

Here’s the views on New Savanna over the past three months. Very spiky.

Here’s the last year; notice the overall spikiness and the heavy action in August and September.

This is how things look over the whole life of the blog:

I have no idea what caused all that action in 2017, nor what’s the source of the recent, but highly variable, increase since August, 2023.

Here’s the most popular posts for the last three months (click on chart to embiggen):

Note particularly the most popular one: GOAT Literary Critics: Part 1.1, What do you mean, literary critic? That’s from December 4, 2023, and is the first post in my ongoing series on literary criticism. That series is a response to Tyler Cowen’s recently published book, GOAT: Who is the Greatest Economist of all Time and Why Does it Matter? Cowen published a link to my post at his blog, Marginal Revolution, on December 6 and that drove a lot of traffic to New Savanna that day.

The second post on the list is my post on Ninja Scroll; as I indicated above, it’s the most popular one at New Savanna and dates from September, 2010, the first year of the blog. In third place we have one the posts in my series about Miyazaki’s The Wind Rises. That is now the third most popular post on the blog and dates from late 2015.

Who’s the GOAT of Economics? Tyler Cowen on His New AI Book & More! [Bonus: from double-entry bookkeeping to supply and demand] – that’s in fourth place and has some comments on a podcast in which Cowen discussed his GOAT economists book. During the conversation he mentioned that he couldn’t understand why 17th century thinkers had such trouble conceptualizing the phenomenon of supply-and-demand, which he could easily explain to a reasonably intelligent teenager. I did some thinking about it and made a variety of comments about it, giving special attention to a speculation that the conceptualization probably piggy-backed on the practice of double-entry book-keeping. Cowen apparently liked that and gave me another link.

More later.

Report from Academia.edu • I’m Back! • Mar 22, 2023

Here’s the stats from 7:30 this morning:

I’m definitely down from the peak of March 13, but I seem to be running a bit higher than things were in February. Can I keep it up? Who knows.

But here’s the big news:

I’m now in the 99.9th percentile (top 0.1%) of Academia users. I had been running at the 99.5th percentile (top 0.5%). When you consider the nature of this distribution, that’s a significant jump.

Last March is the first time I got this high in the rankings, and that was an unexplained fluke. I had 875 views on March 28. I have no idea what caused that, but I was unable to sustain it. Perhaps things will go better this time. We’ll see.

The hits just keep coming on my Academia.edu page

I’d earlier reported that a post at LessWrong was driving traffic to my Academia.edu page to look at papers on ChaGPT and ayahuasca. That post – The idea that ChatGPT is simply “predicting” the next word is, at best, misleading – went up on Feb. 20, peeking three days later on the 23rd. That’s when hits began arriving at Academia.edu. Then things slowed down, but never quite back to where they had been before. And there was a smaller spike or two.

But then – SHAZAM! – there was a BIG spike yesterday, March 13. That’s it there at the right:

It’s at 133 document views, 121 unique visitors, while the Feb 23 spike is at 93 document views and 80 unique visitors. What happened there?

Jerry Seinfeld is what happened. I’d posted a new working paper, Seinfeld's Comedy, Jokes are Intricately Crafted Machines, and that got posted at Hacker News (it’s scrolled way down by now). Now, look at this screenshot, which is from my Academia.edu page some time yesterday:

So those “ycombinator.com” links in the 2nd column from the right. That’s the Hacker News. There’s also one from scottaaronson.blog and one from academia.edu.

I wonder where things will go next?

A Small Surprise on my Academia.edu page

I’d expected that the increased activity that started on the 20th would have subsided by now. Things were certainly headed in that direct, but then yesterday, the 27th, they started back up. The cause of that upswing is not obvious. My article at 3 Quarks Daily had some effect. So did my piece above, Faculty psychology and the will. And LessWrong is still sending folks over. 

Where will the numbers go today?

Feb 28 – 5:44 AM – Looking back thirty days

Feb 28 – 5:44 AM – Looking back sixty days

Where things stood on the 24nd.

Where things stood on the 22nd.

Things are still lively on my Academia page

Like I reported the on the 22nd, those spikes at the right reflect a lot of action coming in from LessWrong. People seem to be interested in two papers: the tantalizing future paper about GPT, and a review of a book about Ayahuasca.

Thirty-day activity:

Sixty-day activity:

Another big jump at Academia.edu [LessWrong] – Turn! Turn!

I posted my first working paper about ChatGPT (Discursive Competence in ChatGPT, Part 1: Talking with Dragons) to Academia.edu on Jan. 4, 2023. That leads to the first spike at the left, which is on Jan. 8. I posted my second major working paper (ChatGPT intimates a tantalizing future; its core LLM is organized on multiple levels; and it has broken the idea of thinking.) on Jan. 23, 2023. I am pretty sure that the spike in the center, which is on Jan. 24, is a reflection of that. I posted two minor papers between those two, one on stories, the other on The Towers of Warsaw. The net result was an increase in activity on my Academia.edu page.

60-day Activity at Academia.edu (Feb 22, 2023 – 6:51 AM)

The more dramatic spike at the right, on Feb. 20, has a more interesting cause.

30-day Activity at Academia.edu (Feb 22, 2023 – 6:51 AM)

On February 19 I posted a mid-length piece here at New Savanna, The idea that ChatGPT is simply “predicting” the next word is, at best, misleading. I then cross-posted it to LessWrong. To my surprise and delight, it kicked off a lively conversation over there, which I have found quite useful, BTW. That spike to the right reflects that conversation.

How could that be, you ask, as a post at LessWrong has nothing to do with your Academic.edu page? That’s not quite true. That post contained links to two papers I’d posted there. One is the tantalizing future paper about ChatGPT. The other link goes to a review I did of a book about ayahuasca. Why, you might ask, did I drop such a link into a post about ChatGPT? Because it refereed to some work by the late Walter Freeman, a neuroscientist who studied the complex dynamics of the brain.

So, Academia keeps track of the sources of traffic to papers posted there. When I look at that record for Feb 20, 21, and now 22, I see links to those two papers originating at LessWrong, with more links to the ChatGPT paper than to the ayahuasca paper. That members of the LessWrong community should be interested in the ChatGPT paper is not at all surprising; the community was founded on interest in AI, the possibility of AI Doom in particular. The interest in ayahuasca is not quite so expected. Nor is it unexpected either. Remember, the personal computer revolution got started in Silicon Valley, which is only a stone’s throw from the Haight-Ashbury district of San Francisco. And that was the Mecca of psychedelic hippiedom. And if you read my early autobiographical memoire, Touchstones • Strange Encounters • Strange Poems • the beginning of an intellectual life, you’ll see connections between computing, the idea of recursion, and psychedelia, Coleridge’s “Kubla Khan,” in my own intellectual history.

How does the song go?

To everything turn, turn, turn
There is a season turn, turn, turn
And a time to every purpose under Heaven

A time to be born, a time to die
A time to plant, a time to reap
A time to kill, a time to heal
A time to laugh, a time to weep

To everything turn, turn, turn
There is a season turn, turn, turn
And a time to every purpose under Heaven

A time to build up, a time to break down
A time to dance, a time to mourn
A time to cast away stones
A time to gather stones together

To everything turn, turn, turn
There is a season turn, turn, turn
And a time to every purpose under Heaven

A time of love, a time of hate
A time of war, a time of peace
A time you may embrace
A time to refrain from embracing

To everything turn, turn, turn
There is a season turn, turn, turn
And a time to every purpose under Heaven

A time to gain, a time to lose
A time to rain, a time of sow
A time for love, a time for hate
A time for peace, I swear it's not too late

– Pete Seeger (popularized by The Byrds)

All settled down on my Academia page, Part 3

Back at the end of March I noticed that my view count on my Academica page was blowing up. As you can see from this snapshot I took on May 8, things are now back to normal:

This sixty day shot gives you the whole story:

That was a remarkable spike. Still don't have a clue about what caused it. But then New Savanna itself has had a rather spiky time of it. Here's the whole run, from the beginning in April 2010 to now:

It looks like things have gotten back to normal on my Academia page, Part 2

 I'd previously reported that my view-count at Academia.edu was blowing up:

• What’s going on at my Academia page at the end of March 2022, and
• It looks like things have gotten back to normal on my Academia page

They're still back to normal. I don't expect any deviations from that in the near future.

But I do wonder what happened.

It looks like things have gotten back to normal on my Academia page

Back on April first I’d posted that, for some unknown reason, action on my Academia page had all of a sudden skyrocketed to 875 on March 29, and then quickly dropped down below 400. Would things drop back to the way they’d been previously or stabilize around a higher value, say 80 to 100 hits a day?

Things have now dropped back to the old regime:

I took that image this morning, April 14. The total dropped into the old regime on April 7 and have now remained there for a week. To be sure, I took that shot at 8:06 AM, so there’s plenty of time for today’s number to rise to unexpected heights. But I have no reason to expect that it will.

What’s going on at my Academia page at the end of March 2022

This graph depicts activity on my Academia.edu page from March 2, 2022 to April 1, 2022 (click on the graph to enlarge it):

What happened at the end of March?

The upper green line depicts the number of views for individual documents while the lower black line depicts the number of document downloads. We see a dramatic increase in the number of views. Here’s the number of views at the end-of-day for March 25 through April 1.

March 25: 34
March 26: 61
March 27: 104
March 28: 875
March 29: 388
March 30: 184
March 31: 249
April 1:     44 (7:39 AM)

The total for March 28, 875, is by far the largest number of views I’ve had in a single day. The total for the previous day, 104, may well be the largest number of views up to that date (I started posting at Academia a decade or so ago). The number of views per day is generally between 10 and 50. An older post on activity at my Academia page gives a sense of my normal action at the site: How am I doing at regulating my action at Academia.edu?

Why the sudden increase? I don’t know. However, once I noticed that the increase from the 27th continued on the 28th, I began tweeting links to various papers. I noticed that almost immediately after I’d posted a tweet to my Twitter feed, views of that document would show up on my Academia page. Obviously someone – bots perhaps? – was watching my tweet stream. I note that the increased activity seems to be coming from a number of different geographic locations.

I must have tweeted 20 or 30 or more links on the 28th. The total for the 29th, 388, was down considerably, but still well above the usual. I did some tweeting of papers on the 29th, but not so many. The totals for the 30th and 31st are still well above usual, with some tweeting of papers on my part. Note that I captured the graph at 7:39 AM on April 1, when there were 44 views. I don’t know what the number will be for the end of the day.

How am I doing at regulating my action at Academia.edu?

On Oct. 31 I posted this chart in a post that was otherwise about Donald Trump and his tweeting activity:

The chart depicts action on my page at Academica.edu, my main repository of articles and working papers. The green (upper) line shows how many papers views I got each day while the brown (lower) line shows downloads.

The point of the chart is that there are two periods of markedly increased activity, activity caused by my deliberate and conscious action. I then asked: “What are the chances that I can keep it going? Off hand, not good. I just don’t have that much of what seems to be the right material to keep it going.” That is, I knew what I did to drive the numbers up but I wasn’t prepared to keep doing that.

This chart depicts my action as of November 16 and 10:32 PM:

That certainly looks like I’ve managed to keep the action up for roughly the past month. To be sure, the action is spiky but it does appear that, on average, interest is up.

Here’s a chart I grabbed at 5:52 this morning, November 19, 2018:

We’ve got a new spike there at the right edge of the chart. While it’s low in relation to the action for Nov 4 and 5, the high point of the chart (and I forget just what I did to produce that), it’s high in relation to the chart as a whole and, in particular, it’s in the range of the mid-September action that started this roll (and which is now trailing off the left edge of the graph).

I know exactly what I did this time. I sent the following note to the Humanist Discussion Group:

I don’t know how old you are, Jim (if i may), but I’ll be 71 in a few weeks and was publishing on literature and computation in the mid-1970s:

"Cognitive Networks and Literary Semantics", MLN 91: 1976, 952-982. Here I used a computational model to examine the semantic structure of Shakespeare’s sonnet 129. There’s a downloadable version of that article here: https://www.academia.edu/235111/Cognitive_Networks_and_Literary_Semantics

That same year David Hays and I published this, “Computational Linguistics and the Humanist”, Computers and the Humanities, Vol. 10. 1976, pp. 265-274. There we proposed something we called Prospero, a computer system for reading Shakespeare in some interesting way. Alas, the more we learn about both computing and about the mind/brain, the most distant that Prospero seems, but still, it’s worth thinking about. You can download that here: https://www.academia.edu/1334653/Computational_Linguistics_and_the_Humanist

Best,

Bill Benzon

That recent spike consists mostly of interest in those two papers, which is much stronger on views than actual downloads (as is generally the case).

I have no idea what will happen today. Of course, I do expect the numbers to rise above the floor, after all it’s 6 AM here on the East Coast of the USA, but just how high they’ll go, I still don’t know. Nor will I hazard a prediction about the future. But I’m pretty sure that if I want to keep the numbers up, I’ve got to upload a paper to the “sweet zone” pretty soon, and I don’t have anything that’s quite ready for upload.

We’ll see.