ChatGPT going haywire today in Eutychius

This afternoon I have been working on the end of Eutychius, chapter one, where we encounter the story of Joseph and his less-than-amazing family problems.  As before I have interleaved the sentences: Italian first, my original translation second, and then the output from ChatGPT.

But today it’s not working at all.  Today ChatGPT is generating hallucinations, not just once, but many times.  Here’s one:

S’abbatté sulla Siria una grave carestia. Giacobbe disse quindi ai suoi figli: “Andate in Egitto a comprare del grano” (148). I fratelli di Giuseppe andarono in Egitto.

A severe famine struck Syria.  Jacob then said to his sons, “Go to Egypt and buy grain.”(148)  Joseph’s brothers went to Egypt.

A severe famine spread over the whole country of Syria. Jacob said to his sons, “Why do you just keep looking at each other?” He continued, “I have heard that there is grain in Egypt. Go down there and buy some for us, so that we may live and not die.”  Then ten of Joseph’s brothers went down to buy grain from Egypt.

Um… what?

In other cases it took direct speech and turned it into indirect speech, sometimes omitting material.

Today I have a little time, so I was going to do some work.  But no.  Basically I can’t get anything done today.  The output from ChatGPT translation is switching around madly.

Yesterday I heard of the term “temperature”, used for one of the parameters to AI generated text.  Like so much about AI, the name is intended to obfusticate.  Many of the “explanations” are likewise intended to conceal.  This one is a bit better:

A temperature of 0 means roughly that the model will always select the highest probability word. A higher temperature means that the model might select a word with slightly lower probability, leading to more variation, randomness and creativity.

and here’s another:

TEMPERATURE is the variable in AI systems that determines how predictable or not it is.

HIGH temperature will cause it to be more creative.

LOW temperature will cause it to be more predictable.

High temperature ALSO can cause it to output complete nonsense if the temperature is too high.

Essentially it’s a parameter to say “how closely do you want to follow the data in the database, and how much randomness do you want?”

None of us need “tools” that aren’t there when you need them and vary randomly in output.

Share

A few months of interesting links

For some months I’ve been collecting bits and pieces.  Mostly I have nothing much to add, but they shouldn’t be lost.

Cool 9th century manuscript online as PDF

Via Rick Brannan I learn that a downloadable PDF of the Greek-Latin St Gall 9th century manuscript of Paul’s letters is online and can be downloaded as a single PDF:

Note the link on this page where you can download a PDF of what appears to be the entire Codex Boernerianus. It is beautiful.

And so you can.  It’s at the SLUB in Dresden here, where it has the shelfmark A.145.b.  It also contains Sedulius Scottus, I gather.

Nice to see the interlinear, isn’t it?

Codex Trecensis of Tertullian online

A correspondent advised me that the Codex Trecensis of the works of Tertullian has appeared online in scanned microfilm form at the IRHT.  Rubbish quality, but far better than nothing.  The ms is here.  De Resurrectione Carnis begins on 157r and ends on 194r.  De Baptismo begins on folio 194r and ends on 200v.  De Paenitentia begins on folio 200v.

Saints lives = Christian novels?

A review at BMCR by Elisabeth Schiffer of Stratis Papaioannou, Christian Novels from the ‘Menologion’ of Symeon Metaphrastes. Dumbarton Oaks medieval library, 45. Harvard University Press, 2017, caught my eye.   This contains 6 lives from Metaphrastes collection.

Even though hagiographical texts are among the most frequently translated Byzantine sources, little effort has been made so far to translate parts of Symeon Metaphrastes’ Menologion. This is primarily due to the generally unfortunate editorial situation of these texts: They are transmitted relatively standardized, but in a vast number of liturgical manuscripts.

In addition to summarizing the status of research on Symeon’s rewriting enterprise, Papaioannou explains in his introduction why he calls the texts in focus “Christian novels.” It is not unproblematic to apply this modern term, as he himself states, but he decided to do so because of the fictionality of these narratives and because of their resemblances to the late antique Greek novel. When saying this, it is important to emphasize—as Papaioannou explicitly does—that these texts of novelistic character were not understood as such by their audience. On the contrary, the Byzantines regarded these texts as relating true stories, written for edification and liturgical purposes (see pp. xiv-xviii).

It’s an interesting review of a neglected area of scholarship where the tools for research – editions and translations – are not available.

Full-text of the Greek Sibylline Oracles online for free

Annette Y Reed broke the story on Twitter: it’s J. Geffcken, Die Oracula Sibyllina, Leipzig: Hinrichs, 1902, which has turned up at Archive.org here.   A useful transcription, rather than the original book, is also online here.

All known mss in the Bodleian library – detailed in online catalogue

Ben Albritton on Twitter shares:

This is awesome – “This catalogue provides descriptions of all known Western medieval manuscripts in the Bodleian Library, and of medieval manuscripts in selected Oxford colleges (currently Christ Church).” Sharing ICYMI too.

It also has direct links to the for Greek mss!

Where did the Byzantine text of the New Testament come from?

Peter Gurry at the ETC blog asks the question, and suggests that Westcott and Hort are no longer the authorities to consult.

How to respond to politically motivated persecution

Since the election of President Trump I have noted on Twitter a new form of anti-Christian posting.  There has been an endless stream of anti-Christian jeering online, demanding “how dare you support Trump”?  It is surreal to see how people who hate Christians suddenly have become expert theologians on what Jesus would do.  Thankfully a certain Kurt Schlichter writes *Sigh* No, Being A Christian Does Not Require You Meekly Submit To Leftist Tyranny:

Everyone seems to want to tell Christians that they are obligated to give in. There’s always some IPA-loving hipster who writes video game reviews when he’s not sobbing alone in the dark because no one loves him tweeting “Oh, that’s real Christian!” whenever a conservative fights back. I know that when I need theological clarification, I seek out the militant atheist who thinks Christ was a socialist and believes that the Golden Rule is that Christians are never allowed to never offend anyone.

It’s a good article, and sadly necessary in these horribly politicised times.  It’s worth remembering that, were times different, rightists would most certainly adopt the same lofty lecturing tone.

A quote for pastors from St Augustine

Timothy P. Jones posted on twitter:

“If I fail to show concern for the sheep that strays, the sheep who are strong will think it’s nothing but a joke to stray and to become lost. I do desire outward gains–but I’m more concerned with inward losses” (Augustine of Hippo).

Queried as to the source, he wrote:

It’s from Sermon 46 by Augustine–the entire message is an outstanding exposition of what it means to be a shepherd of God’s people…. I translated the above from thisHere’s a good English translation as well.

Artificial Intelligence in the Vatican Archives

I knew it.  It’s alive!!!

Well, not quite.  This is a piece in the Atlantic, Artificial Intelligence Is Cracking Open the Vatican’s Secret Archives: A new project untangles the handwritten texts in one of the world’s largest historical collections:

That said, the VSA [Vatican Secret Archives] isn’t much use to modern scholars, because it’s so inaccessible. Of those 53 miles, just a few millimeters’ worth of pages have been scanned and made available online. Even fewer pages have been transcribed into computer text and made searchable. If you want to peruse anything else, you have to apply for special access, schlep all the way to Rome, and go through every page by hand.

But a new project could change all that. Known as In Codice Ratio, it uses a combination of artificial intelligence and optical-character-recognition (OCR) software to scour these neglected texts and make their transcripts available for the very first time.

They’ve found a way around the limitations of OCR by using stroke recognition instead of letter recognition.  They open-sourced the manpower by getting students (who didn’t know Latin) to input sample data, and started getting results.

All early days, but … just imagine if we could really read the contents of our archives!

Kazakhstan abandons Cyrillic for Latin-based alphabet

Via SlashDot I read:

The Central Asian nation of Kazakhstan is changing its alphabet from Cyrillic script to the Latin-based style favored by the West. The change, announced on a blustery Tuesday morning in mid-February, was small but significant — and it elicited a big response. The government signed off on a new alphabet, based on a Latin script instead of Kazakhstan’s current use of Cyrillic, in October. But it has faced vocal criticism from the population — a rare occurrence in this nominally democratic country ruled by Nazarbayev’s iron fist for almost three decades. In this first version of the new alphabet, apostrophes were used to depict sounds specific to the Kazakh tongue, prompting critics to call it “ugly.” The second variation, which Kaipiyev liked better, makes use of acute accents above the extra letters. So, for example, the Republic of Kazakhstan, which would in the first version have been Qazaqstan Respy’bli’kasy, is now Qazaqstan Respyblikasy, removing the apostrophes.

The article at SlashDot instinctively opposed a change, which can only benefit every single Kazakhstani, by making a world of literature accessible.  Ataturk did the same, and for the same reason.

Tell Google that a book is in the public domain

Sometimes Google misclassifies books.  But there is a way to tell it that actually the book is public domain.  The Google link is here.  From It’s surprisingly easy to make government records public on Google Books:

While working on a recent story about hate speech spread by telephone in the ’60s and ’70s, I came across an interesting book that had been digitized by Google Books. Unfortunately, while it was a transcript of a Congressional hearing, and therefore should be in the public domain and not subject to copyright, it wasn’t fully accessible through Google’s archive….

But, as it turns out, Google provides a form where anyone can ask that a book scanned as part of Google Books be reviewed to determine if it’s in the public domain. And, despite internet companies sometimes earning a mediocre-at-best reputation for responding to user inquiries about free services, I’m happy to report that Google let me know within a week after filling out the form that the book would now be available for reading and download.

What does it mean to speak of an authorial/original/initial form of a Scriptural writing when faced with tremendous complexity in the actual data itself?

Back at ETC blog, Peter Gurry discusses this with Greg Lanier here.

Some of the difficulty, one senses, is because the interaction of the divine with an imperfect world is always inherently beyond our ability to understand.  It requires revelation, which is not supplied in this case.

And with that, I think I’ve dealt with a bunch of interesting stories which didn’t deserve a separate post.  Onward!

Share