1453 stories
1 follower

BBC Africa Book Club: 'Kintu' by Jennifer Nansubuga Makumbi

1 Comment
#BBCAfricaBookClub reviews Kintu by Jennifer Nansubuga Makumbi, with Princess Irede Abumere and guest Sope Martins.
Read the whole story
5 hours ago
the great Ugandan novel
Share this story

The persistence of memory

1 Share

The scent she wore instantly evoked fornication. I can’t say today whether it was woodsy or floral, musky or fresh. But the images it called to mind were obscenely vivid. I would say pornographic but the imagery transcended mere human sexuality. This was deeper, even more primal. Seed pods puckering wetly open. Proboscises puncturing nectar domes. Tentacles winding toward a sticky crevice. There was something strangely botanical about all of this. It left you feeling like an incredibly horny plant.

Jane was an intern at the magazine where I worked. Fifteen years later, I can still clearly recall her features—though she always gave the impression of someone who wanted to be overlooked. A quiet, demure persona with a crouched posture and a low voice.

As well as disrupting my ability to focus, Jane’s perfume left me struggling with a disturbing sense of sexual trespass, as if my eyes had clicked to a particularly low décolletage. Of course, ogling with your eyes is one thing, but how do you keep from ogling with your nose?

There ought, I thought at the time, to be a convention regulating the use of such perfumes, much as chemical agents are regulated in war. At the very least it seemed like I should have had the right to ask what this scent was. To diminish its power by learning its name. But asking about a woman’s perfume is like asking about a guy’s weave. In polite society it’s just not done.

My only recourse was to start wearing a scent of my own, as a kind of olfactory self-defense. I had never worn cologne. Nor had I ever interpolated French phrases into my speech or affected a silver knobbed cane. I’m an ordinary person. I don’t presume to take up more social space than anyone else. But then I’ve heard it said that within hours of the first use of chemical weapons by the Germans in WWI, a British general was on the line to London, demanding something similar. War, in other words, is war.         

And so the battle commenced. The ribald twang of Jane’s succulent scent versus the citrusy fizz of mine. There was no way to know who was winning, of course. This was a covert war, the outcome never to be known. As far as I was concerned, it was mostly a war between me and my own nostrils, which reared like stallions whenever she was near.

But all this was years ago. And things change, as we know. The eyes cloud over. Dull satiety replaces that ache in the throat. Suffice it to say that there comes a point when the past itself begins to seem seductive, exuding an alluring fragrance of its own. You begin, like Orpheus, to look back.

Returning is never a good idea. For every nostalgic indulgence you always come away feeling measurably diminished. As if you had robbed yourself of something and yet somehow failed to obtain it in the process. Besides, if you ever do reach a point where the past begins to seem more appealing than the present it’s probably the present that needs your attention—not the past.

I’ve been good about not returning, for the most part—no reunions, no Facebook archeology, no (thank god) “dropping in” on old science teachers. But in the hollow of mid-life one is inclined to allow oneself something. An affair is too skeevy. A sports car, too cheesy. I toyed for a while with the idea of remodeling the bathroom. My in-laws have a royal bathroom, with separate stalls on either side of a vast vanity. Imagine! I wouldn’t aspire to anything so grand. 

But this. This small indulgence, I decided, I would allow—as a kind of mid-life gift to myself. The name of Jane’s perfume.

Naturally Jane’s last name was as common as dandelions. The world swarmed with her doppelgängers. Lawyer Janes, professor Janes, city council member Janes. One Jane was a staff assistant to the U.S. Senate Committee on Appropriations. Could there possibly be a career path connecting this gig with a dinky magazine internship?

Switching to Facebook, I shuffled through another deck of Janes, eyeballing the pictures. Perhaps she preferred anonymity. Perhaps she hadn’t changed a bit. Perhaps she was still trying to be overlooked.

A Google image search served up a thousand more strangers. I kept scrolling until the Janes petered out, giving way to the usual internet silt—random film stills, women’s shoes, and high school yearbook pictures of Steven Tyler.

It’s unusual, these days, not to be able to find what one is looking for. From vintage Atari consoles to Brazilian hair extensions, all can be found and swiftly delivered. But Jane could not be found, nor delivered, and I began to feel a grudging respect for that fact.

Eventually I realized I was going to have to be more proactive. So I clawed my way to the back of my mind and unearthed the names of the two other interns who worked with Jane at the time. Surely they couldn’t be more difficult to find.

It was not to be. One was in London, the other Pittsburgh, but both had lost touch with Jane the moment the internship ended.

And so, after days circumventing sleazy ads promising details on the scandalous missteps of anyone you’d care to name (divorce records, sexual offenses, bankruptcies, mug shots), I surrendered and shelled out $20 for one day of access to every known Jane’s background report.

By my reckoning, she should have been 38 or 39, so I focused on those Janes, who numbered three. One played organ at a church in New Jersey. That didn’t sound right. Another was an accountant. Which I hoped wasn’t right. The last was a reporter at a business journal in Texas. Which might have been right, I guess. I mean, you have to end up somewhere in life.

But the Texan Jane, I discovered, in the course of a rather awkward phone call, had left that job several years ago. No one seemed to know where she had gone. The guy who hired her had himself moved on.

And where was he?


I reached him eventually, not hoping for much. And talking with him was confronted with the amazing fact that, despite the details we were able to share, neither of us could be certain whether our Janes were the same.

“It’d be interesting to know why this perfume is so important to you,” my shrink said.

 My shrink. A woman I pay to be a pain in my ass. But: fair question.

Be honest, I asked myself. Why is it so important to you? Don’t pretend it’s just idle curiosity. The times arrives when one no longer falls so easily for one’s own little ruses.

Could it have something to do with your waning virility?

Wait, waning virility? Who said anything about waning? 

Okay, let’s say it is about the waning. I wouldn’t put it that way, exactly, however. I wouldn’t have said waning.

What would you say?

Well, maybe that in middle age desire is no longer experienced, as it were, first-hand, but as a kind of bodily rumor. You can make up for this by paying more attention, but it will never be as it was. The tannins of age dim the senses.

So maybe the perfume promised an exception. An electric reminder of what desire feels like in its rawest form.

Certainly it had little to do with Jane herself. Jane never said anything particularly memorable. Nor was there a physical attraction. It was hard to be physically attracted to someone so self-effacing. Even at parties Jane kept to the corners—prim, tidy, self-contained.

Then again I suppose the same could be said of a late summer blackberry. Tidy on the outside. And on the inside: an orgy.

Perhaps what gripped me was the specificity of what I sought. A single brand name, like the name of a movie or song which, at last remembered, bestows a disproportionate peace.

Several weeks later I was wandering the mall and, passing a perfume kiosk, decided to stop and chat. The proprietor, a bald, 60ish man wearing a Ralph Lauren windbreaker and a snowy soul patch, was sitting in a rolling chair, stabbing a jumbo calculator.

Somehow I could tell that he was more of a merchant than a connoisseur of scents. Perhaps it was his business name. “Perfume Guy.”

Indeed, Perfume Guy soon informed me of his impotent nose. Twenty-five years peddling perfume at the mall will do that to you, I guess. But if he was bothered by the loss he didn’t show it.

“There’s a lot of bad smells in this world so sometimes it’s good not to have a sense of smell,” he observed.

I nodded sagely, but was beginning to suspect that Perfume Guy might not have the answers I sought.

But the encounter with Perfume Guy had rekindled my curiosity, and got me thinking that, even if he didn’t have the answers, surely someone must.

It was from Macy’s that I originally bought my defense cologne, 15 years ago, so that was where I headed next.

At first sight, I could tell that Joann was exactly who I was looking for. The opposite of Perfume Guy. Not quite grandmotherly, she looked like someone who had moved past raw physical passion, but was still ready to cackle about it on someone’s front porch.

Joann had warm brown eyes and brittle, shoulder-length hair. A thin cardigan hung from her shoulders, cat glasses from her neck.

She didn’t bat an eye on hearing my problem, but immediately set to work lining up sniff tests. Black Opium was her first thought. But then, Black Opium wasn’t around 15 years ago, so we moved on to something else. Eventually Joann produced a glass jar of coffee beans for me to sniff so I wouldn’t go “nose blind”.

“How old was the woman you smelled it on?” she inquired.

I told her and she gave a musky sigh.

“Well I’m gonna guess that that fragrance doesn’t exist anymore. Because younger people tend to wear trendier fragrances, and those don’t last.”

No, of course not. Not only does Jane elude me, and the name of her perfume, but the perfume itself disappears. I should have taken some hint from this. A reminder of my own lesson. Never look back.

And yet, Chloe’s “Love Story” produced an odd falling feeling in my chest. At the limits of my senses the ghostly outline of a woman appeared. A template of desire. Someone I longed for but could not quite apprehend. It was then that I marked the similarity of that feeling to terror.

Some things simply elude us. Elusiveness, in fact, is the essence of perfume, just as perfume is supposed to be the essence of something else. It’s this regressive withdrawal that keeps you drifting after it. Aching to locate its origin.

I once knew a woman who wore a minty perfume. Her name was Anouk. Anouk’s perfume was intriguing. It raised a question, fomented confusion. A perfume is something that does not quite add up. A puzzle with one key piece cruelly withheld. And what else can one conclude but that the key piece is the woman herself?

The one time I really fell for a girl, I mean really fell, with a kind of witless abandon that I cannot even imagine today, it was because the girl herself resembled a fragrance. Evocative but ephemeral. Nymph-like, always receding (tactically, I realized, too late) as I approached. I was like a dog biting air.

Perfumes are a way to correlate oneself with an incorporeal beauty. Is this an illusion? A mere parlor trick? We are all, after all, little more than flesh and hair.

But if it is an illusion then why does it work so well? Why should we be aroused by the smell of a flower? It’s not like we fucked flowers, eons ago.

Perhaps there is a natural mystery to people that a fragrance can sometimes express. The same way a song can bring on a swoon.

Is there some logic to whether a given fragrance bonds with a person? Because not just any scent will do. It needs to fit. There are trashy perfumes and old lady perfumes. Elfin perfumes and voluptuous perfumes. Surely part of the thrill is finding the one that fits—like a witch finding her familiar. Owl, spider, raven, cat—like perfume, the job of the familiar is to always accompany you, represent you to yourself, help you feel less lonely.

But if this is so, then what could explain why Jane’s perfume was so wickedly unchaste?

“There’s totally a correlation,” Sue Phillips was saying. Between people and the scents they wear, she meant. “Someone who is very sophisticated, who is very sensual, very romantic, is going to want to go for lovely, warm, smooth, sensual, deeper, sexier notes, as opposed to something light, crisp and airy.”

Sue was a perfume industry veteran who seven years ago started her own company specializing in custom scents. She had designed scents for Jamie Foxx, Katie Holmes, and Snooki. These credentials should suffice.

I reached out to Sue by phone because why not? This is the great thing about being a journalist person. If you ever happen to have a question about something you can call up whomever and get an answer.

I liked Sue immediately. She spoke in long lyrical strings of adjectives. She had a great South African accent and used the word lovely a lot. And every now and then she would (as I imagined it) tip back her predacious head and give a great, mannish laugh, and I could picture her broad lipsticked mouth and the way she might squeeze my thigh with a bony hand.

It was Sue’s view that fragrances not only correlate with personality but also reflect something of our past—principally our memories of our parents. Older, more mature women, Sue said, tend to prefer the heavier, floral scents, like Chanel N°5, because that’s what their mothers wore. Men often prefer an “Old Spice kind of citrusy, sporty feel,” because it reminds them of their fathers.

One of the most popular ingredients, Sue said, was vanilla.

“Because it hearkens back to memories of your childhood, and for whatever reason—you’re a good kid, you hurt yourself, or whatever—you get rewarded, and you get a vanilla cookie, or vanilla ice cream, or a vanilla lollipop. And all those wonderful associations make you feel good… It’s a very erotic and exotic aroma.”

I listened with growing alarm. Certainly you expect your shrink to walk you into a Freudian corner now and then—but not a perfume consultant.

But when I name-checked the Viennese maestro Sue did not blanch.

“Yes, I would definitely say that,” she said, and went on to share a borderline pervy story about a prominent Maryland pastor who’d recently come to see her. The pastor’s request was that Sue design a perfume for his wife that smelled of ice cream and donuts, which for him carried fond associations from childhood.

“Isn’t it amazing?” Sue said. “So I created a very beautiful, luscious, edible, fruity fragrance with some lovely vanilla notes.”

“And did he like it?” I inquired.

“It drove him nuts,” Sue said, and gave another huge, head-rocking laugh.

Could it be that Jane’s scent, whatever it was, by the sheerest accident, linked to some buried Oedipal kink? (Never look back!)

It was probably something fruity, light, sweet, Sue said.

“Fruity notes are very edible and very sensual, very luscious. And I’m not talking about apples and oranges. I’m talking about some very delectable, luscious, sort of peachy notes. Peachy notes are very sensual. Peach, and I would even say pomegranate, some of the honeydew melons, to give it a very lovely smoothness and a sexiness without that strident acidity.”

Reader, I’ll admit it. I was aroused.  

With Sue’s rich insights I decided to abandon my search for Jane. I had after all explored every avenue. Made no end of phone calls. I even reached out to the organ lady. There was nothing left to do but invite new puzzles into my brain in hopes of displacing the one I’d never solve.

But then, unexpectedly, one of the lines I’d put out went taut. The Tennessee connection had mentioned the name of a guy he vaguely remembered his Jane having married. I had located a number of men with that name and sent out blind emails. (Excuse me sir, can I smell your wife?)

My inbox dinged, and there was Jane. She had taken her husband’s name, of course, which was why she’d been so hard to find. They were living in Austin, Texas, her husband a professor at UT.

I decided to delay sharing with Jane the nature of my interest. I would tell her in person, if I ever got the chance.

That chance came a few months later, when I was in Austin to see family. Jane and I arranged to meet for coffee.

I arrived early and claimed a lozenge-sized table on a cement balcony overlooking a strip mall parking lot. In the shade the heat was tolerable. Grackles bounced around on black wire legs.

Jane arrived looking perplexed, bemused. In a white, thigh-length sweater and gray leggings she projected casual chic. She had lightened her hair, which sat curled on one shoulder like a well-behaved cat. Conservative earrings, splendid teeth. Her whole aspect seemed far more direct than it had been 15 years earlier—which was only to be expected.

She excused herself to get some water, which left me a moment to collect myself. And to notice that my nostrils were behaving themselves. No wanton fragrance pervaded the air.

A moment later she returned. It was an odd, off-kilter conversation. We took turns filling in the historical blanks, sharing the pretense that our brief work relationship warranted the reunion, which it did not.

After the magazine, Jane had wanted to go to grad school, but her parents insisted she do something practical. So she went to law school, and hated it. Yet she never once thought of dropping out, and when I asked her why she seemed genuinely unable to answer.

“I should have,” she said, after a long pause. “I should have, I just… I didn’t really… it wasn’t something that even occurred to me, to drop out. But I should have…”

I was interested in the time she took to consider this question. The indifference with which she abstracted herself from the conversation. For several moments we just sat there, listening to the empty rush of cars on Lamar Boulevard.

The rest of the story eventually emerged. Her husband specialized in machine learning and abstract math. (“He’s very left-brained.”) Jane met him in law school, and followed him through two post-docs. Then he landed the job at UT, and she found work freelancing for several periodicals, including that business journal. That stopped when the kids came, of course. Three of them, the youngest two years old. She had never practiced law.

Throughout the conversation Jane remained preoccupied, a little hieroglyph of trouble etched on her brow, as if she were struggling to arrange these great, unwieldy blocks of thought in her mind. The nature of the puzzle was unclear to me, at first. Whether it was me, my peculiar presence. Or whether it was herself, a woman who had disappeared into her own life and then briefly surfaced from it on this balcony among the grackles, to be asked absurd questions by a near stranger with unresolved issues of his own.

I began to feel bad. Like I had woken her from something. Never look back, was the rule. Yet that was exactly what I was asking her to do.

What else did she regret? Was this what she was wrestling with? I could have asked about her husband, whose work was so abstruse that he could never talk to her about it. Or her children. Was there a rebellion somewhere that should have happened but didn’t?

But at a certain point life becomes too big to regret. At a certain point, you can no more ask about it than you would ask about a woman’s perfume.

I limited myself instead to asking if it was sadness I was sensing in her involved silences. But Jane only laughed, brightening suddenly, and denied it. She just wasn’t used to this sort of conversation, she said.

We parted soon after. Halfway to the parking lot Jane stopped short and looked at her hand, which still held the water glass. She stared at it in wry confusion. One more puzzle to account for.

She made to return it but I stopped her.

“Never mind,” I said. “Just take it. It doesn’t matter.”

“Why do you say that?” she said, that hieroglyph of trouble appearing again.

“It doesn’t matter,” I said again, this time with greater certainty. “It just doesn’t. None of the things that we tend to think matter actually matter at all.”

She had denied ever having worn perfume, of course. Somehow I wasn’t surprised. Mysteries enjoy themselves too much to disrobe so obligingly.

“It was just never part of my repertoire,” Jane said. Her husband, in fact, was hypersensitive to most fragrances and had asked her not to wear them.

She did mention that she had once used a product called Fuzzy Peach, from the Body Shop. But that was in high school, well before the magazine work.

Or so she thought.

It was far easier finding the Fuzzy Peach than it had been to find Jane. As Joann had foretold, it was no longer being produced. But one could buy a 15ml vial on Ebay for $30.

It arrived one day in a padded envelope. Throat constricting, I tore through the bubble wrap and held up the tiny glass vial, about half the size of my pinky. The glass was clear, as was the liquid within. It was interesting to imagine that herein was contained, in its most concentrated form, the essence of all I was looking for. All I had lost, or never attained in the first place. A seductive notion. That you could capture something, hold onto it. Encase it in glass. Find it on Ebay.

There was no need to open the vial. Even without opening it I caught the scent of peaches. Lovely, luscious. Redolent, I guess, of the high school Jane. Not the one I remembered.

Read the whole story
1 day ago
Share this story

The 6 year old runs up triumphantly, brandishing - huh? "What are these?" Curses silently. "Condoms" Gently takes them out of his hand "But, but what do you use them for?", he demands. Deep breath. "Protection." "Okay". Satisfied, he turns and skips away happily to play.

1 Share

The 6 year old runs up triumphantly, brandishing - huh?
"What are these?"
Curses silently.
Gently takes them out of his hand
"But, but what do you use them for?", he demands.
Deep breath.
Satisfied, he turns and skips away happily to play.

1 like
Read the whole story
2 days ago
Share this story

Event: ‘Abina and the Important Men’, SOAS (21 October)

1 Share

‘Abina and the Important Men’
21 October, 1700-1900
SOAS, London

Abina and the Important Men (2017) was directed by Soumyaa Kapil. It is a lightly animated film adaptation of the award-winning graphic history by Trevor R. Getz and Liz Clarke.  It tells the story of Abina Mansah, a very real 19th century women in West Africa who challenged both the colonial system and institutions of enslavement.  The film is the product of a community of students and scholars from the departments of Africana Studies, Theater, Cinema, and History at San Francisco State University.


Image courtesy of SOAS

Watch the trailer for Abina and the Important Men .

Soumyaa Kapil is a director and producer, as well as the director of the DocFilm Institute at San Francisco State University. Her films regularly engage with issues of the human condition as well as how political landscapes shape identity and power structures within marginalized communities. Amongst her projects: Veteran Documentary CorpsMadame MarsThe Love BoatThe Life of a String Quartet and Abina and the Important Men.

The screening on 21 October will be followed by a Q&A with Trevor R. Getz. Getz is an historian of Africa and his interests include interdisciplinary methodologies, critical theory, and popular ways of thinking about the past. Find out more about Trevor Getz here: https://trevorgetz.org/  

More information about the event at SOAS can be found here.
Contact email: cas@soas.ac.uk

Date: 21 October at 5:00 PM
Venue: Russell Square, Djam Lecture Theatre (DLT), London

Read the whole story
2 days ago
Share this story

Tenure announcement

1 Share
James Mickens in full effect
Read the whole story
4 days ago
Share this story

Danger: Demo!

1 Share

John Seabrook, "The Next Word: Where will predictive text take us?", The New Yorker 10/14/2019:

At the end of every section in this article, you can read the text that an artificial intelligence predicted would come next.

I glanced down at my left thumb, still resting on the Tab key. What have I done? Had my computer become my co-writer? That's one small step forward for artificial intelligence, but was it also one step backward for my own?

The skin prickled on the back of my neck, an involuntary reaction to what roboticists call the "uncanny valley"—the space between flesh and blood and a too-human machine.

The "artificial intelligence" in question is GPT-2, created early this year by Open AI:

Our model, called GPT-2 (a successor to GPT), was trained simply to predict the next word in 40GB of Internet text. Due to our concerns about malicious applications of the technology, we are not releasing the trained model. As an experiment in responsible disclosure, we are instead releasing a much smaller model for researchers to experiment with, as well as a technical paper.

GPT-2 is a large transformer-based language model with 1.5 billion parameters, trained on a dataset of 8 million web pages. GPT-2 is trained with a simple objective: predict the next word, given all of the previous words within some text. The diversity of the dataset causes this simple goal to contain naturally occurring demonstrations of many tasks across diverse domains. GPT-2 is a direct scale-up of GPT, with more than 10X the parameters and trained on more than 10X the amount of data.

Given the organization name "Open AI", it's ironic that the cited result is actually closed — although the recipe is clear enough that anyone with the right skills and a moderate amount of computational resources should be able to replicate the system.

Seabrook's article gives an impressive demonstration of the technology — at nine points in (the online version of) his article (if I've counted right), there's a button labelled "Read Predicted Text". Here's the first one:

If you click on the button, you see what (a version of?) GPT-2 creates as a continuation of what Seabrook wrote up to that point — in this case it's

The other examples are equally spooky.

Later in the article, Seabrook discusses and exemplifies "fine tuning", i.e. adaptation of the GPT-2 system:

GPT-2 was trained to write from a forty-gigabyte data set of articles that people had posted links to on Reddit and which other Reddit users had upvoted. […]

GPT-2 was trained to write from a forty-gigabyte data set of articles that people had posted links to on Reddit and which other Reddit users had upvoted. […]

Yes, but could GPT-2 write a New Yorker article? That was my solipsistic response on hearing of the artificial author's doomsday potential. What if OpenAI fine-tuned GPT-2 on The New Yorker's digital archive (please, don't call it a "data set")—millions of polished and fact-checked words, many written by masters of the literary art. Could the machine learn to write well enough for The New Yorker? Could it write this article for me? The fate of civilization may not hang on the answer to that question, but mine might.

I raised the idea with OpenAI. Greg Brockman, the C.T.O., offered to fine-tune the full-strength version of GPT-2 with the magazine's archive. He promised to use the archive only for the purposes of this experiment. The corpus employed for the fine-tuning included all nonfiction work published since 2007 (but no fiction, poetry, or cartoons), along with some digitized classics going back to the nineteen-sixties.

I presume, though Seabrook doesn't tell us, that the "Read Predicted Text" examples in his article were generated by this "fine-tuned" version of GPT-2.

He gives another impressive example, of GPT-2 continuing a 1950 article that was not part of its training set:

As Seabrook observes, there are some issues:

On first reading this passage, my brain ignored what A.I. researchers call "world-modelling failures"—the tiny cow and the puddle of red gravy. Because I had never encountered a prose-writing machine even remotely this fluent before, my brain made an assumption—any human capable of writing this well would know that cows aren't tiny and red gravy doesn't puddle in people's yards. 

And he notes that such errors are in fact pervasive:

Each time I clicked the refresh button, the prose that the machine generated became more random; after three or four tries, the writing had drifted far from the original prompt. I found that by adjusting the slider to limit the amount of text GPT-2 generated, and then generating again so that it used the language it had just produced, the writing stayed on topic a bit longer, but it, too, soon devolved into gibberish, in a way that reminded me of hal, the superintelligent computer in "2001: A Space Odyssey," when the astronauts begin to disconnect its mainframe-size artificial brain.

This undermines the claim of GPT-2's creators that they've withheld the code because they're afraid it would be used to create malicious textual deep fakes — I suspect that such things would routinely exhibit significant "world modelling failures".

And Seabrook's experience illustrates a crucial lesson that AI researchers learned 40 or 50 years ago: "evaluation by demonstration" is a recipe for what John Pierce called glamor and (self-) deceit ("Whither Speech Recognition", JASA 1969). Why? Because we humans are prone to over-generalizing and anthropomorphizing the behavior of machines; and because someone who wants to show how good a system is will choose successful examples and discard failures. I'd be surprised if Seabrook didn't do a bit of this in creating and selecting his "Read Predicted Text" examples.

In general,  anecdotal experiences are not a reliable basis for evaluating scientific or technological progress; and badly-designed experiments are if anything worse. For a bit more on this, from my perspective as of 2015, see these slides for a talk I gave at the Centre Cournot.

This is not deny that GPT-2 and its ilk are impressive, interesting, and useful developments. And if you want to start to understand how and why this all works, see Andrej Karpathy, "The Unreasonable Effectiveness of Recurrent Neural Networks" (May 21, 2015), which illustrates the predictive-text capabilities of a much smaller and simpler "deep learning" system, with code that you can run yourself on a modern laptop.

But I don't think that John Seabrook needs to worry about being replaced by GPT-2 — or even GPT-3 or GPT-4. For one of many reasons, read about the Winograd Schema Challenge, a type of text-based problem that tests exactly the abilities needed to avoid "world modelling failures" — and which the best AI systems so far don't do well on.



Read the whole story
7 days ago
Share this story
Next Page of Stories