Question About the Synoptic Problem and the Priority of Matthew

[Swammerdami]

(Or was Matthew FIRST, and Gadarenes an error that Mark/Luke knew to correct?)

This makes no sense. I'd cross it out but the Edit window has expired.

 

[Tigers!]

I’m sorry to be so long getting back here, but real life intervened. So what I’ll do today is give a brief outline of the arguments for the existence of Q.

My author, Burton Mack, whom I referenced in the OP, devotes more than a chapter to the nature and history of Q arguments, and I won’t try to summarize it all here. Instead I will reference an easily accessible web page maintained by one Stephen C Carlson, who Google tells me is an associate professor of Biblical studies at the Catholic University of Australia.

The existence of Q is theorized as a consequence of the Tw0-Source Hypothesis (2SH). Carlson summarizes the argument in a nutshell:

Mark was a source for Matthew and Luke, both of whom also independently used a now lost sayings source called Q.

There are two lynchpins to the 2HS. The first is the primacy of Mark, that is, Mark was the first of the synoptic gospels to be written and Mathew and Luke used it as a source. The second deals with the fact that both Matthew and Luke contain material that is quite similar if not identical, but that is not found in Mark. It is hypothesized that this material, consisting largely of Jesus’ sayings and parables, was found in a second source, now lost, which Matthew and Luke independently made use of.

I will now expand on these two lynchpins.

Lynchpin One: The Primacy of Mark

Carlson identifies two main classes of argument:

1. Arguments from Content

2. Arguments from Wording

Arguments from Content

Quoting Carlson:

It is argued that it is easier to understand certain material (infancy accounts, Sermon on the Mount) being added to Mark by Matthew and Luke than the reverse of Mark's omitting them from Matthew and Luke. It is also argued that it is easier to view Matthew's and Luke's relative brevity in the account all three share as both Matthew's and Luke's compressing the text of Mark to add their own material rather than Mark's abridging the content and expanding the words of one or both of the others. Furthermore, critics have argued put forth reasons for the specific divergences of Matthew's and Luke's order of their material from that of Mark's as more plausible than the reverse. (Stein 1987: 48-51; Tuckett 1992: 264-265)

Arguments from Wording

Again quoting Carlson:

At a general level, proponents of Markan priority find Mark's less literary diction, grammar, redundancy, difficulty of expression, Christology, and use of Aramaic to more likely be intentional improvements by Matthew and Luke rather than Mark's "dumbing down" of one or both of the others (Stein 1987:52-67; Tuckett 1992: 265-267). On a more specific level, Markan priority is also found to in instances where Matthew and Luke seem to occasionally refer to omitted explanatory material in Mark, in Matthew's adding his own theological emphases rather in than Mark's removing them, and an uneven distribution of Mark's stylistic features in Matthew. (Stein 1987: 70-83)

(Incidentally, Carlson references your guy Farrer as a “notable” theorist who accepts the priority of Mark but disparages the existence of Q.)

Lynchpin Two: The Existence of Q

Carlson’s brief summary:

The existence of Q follows from the conclusion that Luke and Matthew are independent … Therefore, the lite[r]ary connection … must be explained by an indirect relationship, namely, through use of a common source or sources.

Again, Carlson identifies two main classes of arguments:

1. Arguments from Content and Argument

2. Arguments from Wording

Arguments from Content and Argument

The assumption that Luke did not use Matthew as a source, and vice-versa, is indicated by the fact neither copies the other’s non-Markan, non-sayings (Q) material. Furthermore, apparently “neither knew the other's arrangement of the non-Markan material with respect to the Markan outline.”

Arguments from Wording

The Q hypothesis is used to explain why the form of the material sometimes appears more primitive in Matthew but other times more primitive in Luke in the face of very impressive exactness in wording in other parts … (e.g. Matt 6:24 = Luke 16:13 for 27 of 28 Greek words and Matt 7:7-8 = Luke 11:9-10 for 24 of 24 Greek words).

One final note, which Carlson doesn’t mention, is that the discovery of The Gospel of Thomas in 1945 proves that “sayings” gospels, as proposed by the Q theory, were a real thing, although so far only Thomas has been discovered.

So those are the bare bones of the Q argument. I have eliminated references to “the double tradition” and “the triple tradition” that Carlson incorporates as I didn’t see that they added clarity to the discussion.

Thank you for that. I had heard of Carlson but not really read much. I will go and see.

 

[Swammerdami]

The following post -- rendering ChatGPT's "opinion" on Gospel dating -- will be "old hat" to experts; and they should ignore this. I include this ChatGPT summary for ignorant laymen like myself.

How certain are scholars about the dating of the Gospels? Paul writes very little about Jesus' life or sermons. Is that because there were no writings available on those topics? Or did the availability of such make Paul feel his efforts would be redundant?

It seems odd to me that Christianity could have blossomed as rapidly as it did without written accounts of Jesus' ministry and sermons. Is it plausible that there was some pre-70 proto-Gospel which was heavily edited into Mark (or Matthew) after 70? ChatGPT views this oppositely:
"Many scholars believe Mark used oral and written sources, including early passion narratives. This would require some passage of time after Jesus’ death (~30 AD) to develop.

And indeed it sounds like even supporters of a pre-70 date place Mark's Gospel after 60 AD.

What is the evidence that Gospel of Mark was written AFTER 70 AD?

ChatGPT said:

The dating of the Gospel of Mark is a central topic in biblical scholarship. While some argue for an earlier date, most scholars date Mark after 70 AD, primarily due to internal evidence suggesting knowledge of the destruction of the Jerusalem Temple in that year. Here's a breakdown of the key evidence:
1. The Prediction of the Temple’s Destruction (Mark 13:1–2)

“Do you see these great buildings? Not one stone will be left upon another; all will be thrown down.” (Mark 13:2)

Also in Matthew and Luke:

And as some spake of the temple, how it was adorned with goodly stones and gifts, he said, As for these things which ye behold, the days will come, in the which there shall not be left one stone upon another, that shall not be thrown down. And they asked him, saying, Master, but when shall these things be? and what sign will there be when these things shall come to pass? And he said, Take heed that ye be not deceived: for many shall come in my name, saying, I am Christ; and the time draweth near: go ye not therefore after them. But when ye shall hear of wars and commotions, be not terrified: for these things must first come to pass; but the end is not by and by. Then said he unto them, Nation shall rise against nation, and kingdom against kingdom: And great earthquakes shall be in divers places, and famines, and pestilences; and fearful sights and great signs shall there be from heaven. But before all these, they shall lay their hands on you, and persecute you, delivering you up to the synagogues, and into prisons, being brought before kings and rulers for my name's sake.

This verse strongly resembles the actual destruction of the Temple by the Romans in 70 AD during the First Jewish-Roman War. Many scholars argue that:

The level of detail and tone suggest that Mark is not just recording a prophecy of Jesus, but is writing after the event and placing it retrospectively into Jesus’ mouth.

Apocalyptic language in Mark 13 is consistent with post-70 AD Jewish and Christian writings processing the trauma of the Temple’s destruction.

Scholarly consensus: This is the strongest internal clue suggesting a post-70 AD date.

2. Lack of Urgency About the Imminent Return of Jesus

Mark seems to reflect a period where the early Christian community had begun to grapple with the delay of the Parousia (Second Coming of Christ).

Earlier sources like Paul’s letters (e.g. 1 Thessalonians) show more immediacy about the return of Jesus.

Mark, by contrast, presents a longer-term perspective, including the suffering of the community and the rise of false prophets (Mark 13:5–13).

This shift suggests that some time had passed since Jesus' death — likely post-70 AD.
3. Dependence on Earlier Traditions

Many scholars believe Mark used oral and written sources, including early passion narratives. This would require some passage of time after Jesus’ death (~30 AD) to develop.

Dating Mark after 70 AD allows time for:

Theological development

Formation and circulation of early Christian traditions

Community reflection on major events like the Temple’s destruction

4. Contextual Clues from Roman Persecution

Some see hints of Roman persecution (e.g. Mark 8:34–38; 13:9–13) that might reflect the Neronian persecution (circa 64 AD) or general post-war hostility toward Jews and Christians.

This may suggest:

A Roman setting (perhaps Rome itself)

A Christian community under stress

A date after or during these traumatic events

5. Patristic Testimony

Early Church Fathers (e.g. Irenaeus, ~180 AD) state that Mark wrote his Gospel after the deaths of Peter and Paul — which tradition places in the mid-60s AD, though this is less precise evidence than internal textual clues.
Summary of Evidence for a Post-70 AD Date
Evidence Description Implication
Mark 13 Destruction of the Temple described in detail Likely written after 70 AD
Delayed Parousia Community prepared for long-term suffering Reflects post-Temple realism
Developed tradition Use of earlier sources and theology Requires time after Jesus’ death
Persecution references Hints of Roman oppression Possibly post-64 AD
Church tradition Mark wrote after Peter’s death Supports post-65 AD date
Scholarly Range

Common scholarly dating: 70–75 AD

Minority early dating: Some evangelical or conservative scholars date it around 50s–60s AD, interpreting the Temple prediction as a genuine prophecy rather than post-event.

Would you like a summary of the case for an earlier date as well?
You said:
Yes.
ChatGPT said:

Certainly. While the majority of scholars date the Gospel of Mark after 70 AD, some argue for an earlier date, typically in the 50s or early 60s AD, before the destruction of the Jerusalem Temple. Here’s a summary of the main arguments for an earlier date:
Arguments for an Early Date of Mark (pre-70 AD)
1. The Temple Prediction Doesn’t Require Post-70 AD Knowledge

Mark 13:2 predicts the destruction of the Temple, but:

Prophetic predictions of the Temple’s downfall were not new — they appear in the Old Testament (e.g., Jeremiah, Micah) and in other Second Temple Jewish writings.

Jesus, as a prophetic figure, may have made a genuine prediction (regardless of its accuracy).

The text does not describe the destruction in vivid historical detail (e.g., no mention of Roman generals, fire, famine, or siege), which one might expect if it were written afterward.

Argument: The prediction could be a true prophecy made before 70 AD, not a post-event fabrication.

2. No Clear Reference to the Actual Event of 70 AD

While Mark 13 contains apocalyptic warnings, nothing explicitly describes the actual events of 70 AD, such as:

The Roman siege of Jerusalem

The burning of the Temple

The fall of specific Jewish leaders

Argument: If Mark was writing after the destruction, why is there no explicit mention of it?

3. Language and Urgency Suggest an Imminent Parousia

Some argue that Mark expects Jesus' return soon, consistent with earlier Christian beliefs:

Mark 9:1 – “Some standing here will not taste death before they see the kingdom of God come with power.”

Mark 13:30 – “This generation will not pass away until all these things take place.”

Argument: This suggests that Mark is writing before a major disconfirmation of Jesus’ return (i.e., before 70 AD and the need to explain a delay).

4. No Mention of Paul’s Letters or Later Christian Developments

Mark shows no awareness of Paul’s theology, church hierarchy, or other developments found in later NT texts (e.g., Matthew, Luke, Acts).

Argument: Suggests Mark was written before these traditions were fully formed and circulated, possibly even before Paul's death (~64 AD).

5. Church Tradition on Mark’s Association with Peter

According to Papias (early 2nd century), Mark wrote down Peter’s teachings:

If Mark was Peter’s interpreter (as Papias says), and Peter died around 64–65 AD in Rome, then:

Mark may have written before or shortly after Peter’s death, possibly in the early 60s AD.

Some place Mark in Rome during Nero’s reign (before the persecution intensified).

Argument: Early dating fits with this traditional account.

6. No Evidence Mark Used the Jewish War as a Source

The Gospel does not reflect Josephus’s account or other details of the war, unlike Luke and Matthew (written later and with more developed responses to 70 AD).

Argument: Mark seems unaware of post-war developments, suggesting an earlier time frame.

Summary of Early-Dating View
Evidence Description Implication
Temple prediction is vague Could be genuine prophecy Doesn’t prove post-70 authorship
Lack of war details No explicit mention of 70 AD events Suggests pre-70 writing
Urgency of Jesus’ return No theological adjustment yet Fits early Christian eschatology
No Pauline influence Simpler theology Possibly before Paul's prominence
Church tradition Mark wrote Peter’s testimony Supports early 60s AD dating
Common Early-Date Proposals

50s–early 60s AD (often around 60–65 AD)

Advocated by some evangelical, Catholic, and conservative scholars, including:

John A.T. Robinson

F.F. Bruce

William Lane

Final Notes

Whether you prefer a pre- or post-70 AD date hinges on how you interpret Mark 13:

Predictive prophecy → likely early date

Post-event writing → likely late date

Would you like a side-by-side comparison chart of the early vs. late dating arguments?

Did Matthew's Gospel borrow from Mark's or vice versa?

ChatGPT said:

Excellent question. The vast majority of modern scholars agree that the Gospel of Matthew borrowed from Mark, not the other way around. This is a key element of what’s known as the “Two-Source Hypothesis,” the most widely accepted solution to the Synoptic Problem (the question of how Matthew, Mark, and Luke are literarily related).
What’s the Synoptic Problem?

Matthew, Mark, and Luke share a great deal of content, often with similar wording, order, and structure.

This suggests literary dependence—that at least some of the Gospel writers used others as sources.

Evidence That Matthew Used Mark
1. Content Overlap

About 90% of Mark is found in Matthew.

When Mark and Matthew tell the same story, Matthew often adds to or expands Mark, but rarely contradicts or omits major episodes (except selectively).

Example: The feeding of the 5,000 (Mark 6:30–44 / Matt 14:13–21) is almost word-for-word in many places, but Matthew smooths out grammar and adds detail.

2. Order of Events

When all three Synoptic Gospels (Mark, Matthew, and Luke) tell the same story, Mark’s order is usually followed by Matthew and Luke.

When Matthew and Luke disagree, Mark usually agrees with one of them, suggesting he’s the earlier “middle term” they both used.

This is a major clue: if Matthew had been written first, Mark would likely not consistently follow its structure.

3. More Primitive Style in Mark

Mark’s writing is rougher, more vivid, and grammatically awkward at times. Matthew smooths out awkward phrases and adds theological refinement.

Mark 1:12 – “The Spirit immediately drove him out into the wilderness”

Matthew 4:1 – “Then Jesus was led up by the Spirit into the wilderness”

Suggests Matthew is editing and polishing Mark’s rougher account, not the reverse.

4. Omissions in Mark of Key Matthew Material

Mark omits major teachings found in Matthew, like the Sermon on the Mount, Lord’s Prayer, and many parables.

If Mark were written after Matthew, it would be hard to explain why he left out so much important material.

Easier to explain Matthew as adding to a shorter, earlier Mark.

5. The “Two-Source Hypothesis”

This is the dominant scholarly theory:

Matthew = Mark + “Q” + special Matthew material (M)
Luke = Mark + “Q” + special Luke material (L)

“Q” (from Quelle, German for “source”) is a hypothetical document that contained sayings of Jesus found in Matthew and Luke but not in Mark.

This helps explain how Matthew and Luke share content not found in Mark, while still both depending on Mark as their narrative base.

Minority Views

Some scholars (especially in earlier centuries) believed Mark abridged Matthew, not the reverse (known as the Griesbach Hypothesis).

This view has largely fallen out of favor due to the evidence above.

Summary
Feature Supports
90% of Mark in Matthew Mark came first
Matthew expands/smooths Mark’s wording Matthew used Mark
Mark’s rougher grammar and style Mark is more primitive
Matthew adds major discourses Mark was a shorter, earlier source
Mark usually preserves the shared order Matthew and Luke drew from Mark
Conclusion:

Matthew almost certainly used Mark as a source, not the other way around. Mark’s Gospel is now widely seen as the earliest of the four Gospels, written around 70 AD, and used by both Matthew and Luke in constructing their own narratives.

Would you like a visual diagram of how the Synoptic Gospels are related according to the Two-Source Hypothesis?

 

[bilby]

ChatGPT is a bullshit generator. Its output is worse than useless, because it is extremely confident and well presented but has absolutely zero reason to be true.

LLMs, such as ChatGPT, take text from a huge variety of undisclosed, unchecked, unverifiable, and therefore utterly unreliable sources, and present "answers" built from the most common components of responses to similar prompts.

It's a bullshit factory. It cares naught for the accuracy or factuality of what it says; It is fundamentaly incapable of testing for such things.

You can get a response from it that sounds convincing. A response that contains all the factors we have learned to use to spot truthful claims, and none of the red flags we have learned to recognise as the indicators of uncertainty, deceit, dishonesty, or ignorance.

As such, its output is perfect for post-truth societies. It's an automated Donald Trump - it lies, not because it seeks to deceive, but just as a matter of habit and routine; It tells the truth, not because it knows or cares what is true, but as an accidental consequence of having spoken at all.

To see it quoted as though its bullshit were in some way worthy of our attention, consideration, and notice makes me both sad and angry.

The very name "Artificial Intelligence" is bullshit.

Please, if you care at all about truth, never use it as though it were a means to generate content in a genuine discussion about reality.

And when you see others use it, please give it the same (or less) weight as a quote from "some guy down the pub said he heard from a mate that his wife's hairdresser thinks..." (Unless it's Gemini, in which case imagine that the guy down the pub also has a swastika tattooed on his face).

 

[Swammerdami]

I have complained about the Age of Disinformation myself. Dangerous for many reasons. The chat-bots and "AI"s are getting better rapidly and I certainly do not see that as an unmixed blessing.

But even worse that relying on LLMs is intellectual dishonesty. The chat-bots are often dishonest -- cf. Elon Musk's Grok -- but some very intelligent humans are themselves unreliable:

ChatGPT is a bullshit generator. Its output is worse than useless, because it is extremely confident and well presented but has absolutely zero reason to be true.

No. ChatGPT does a fair job -- much better than ignorant Swammi could do operating Searches manually even spending an hour or three -- of summarizing what is "out there on the 'Net", perhaps errors and all.

If you had glanced at its output, you'd have seen that it was asked to present BOTH sides of "the controversy." The ChatGPT output would be, at best, a starting point if I were preparing a scholarly article with citations. But it did an excellent job of acquainting the layman with the topic -- or at least the writings that had flooded its LLM. If I had hundreds of hours to spend I might order a dozen books and study them. Lacking that, Google Search, ChatGPT and Wikipedia were my best three choices to get an overview. Of these options, ChatGPT was the best, and its response proves this.

"Proof is in the pudding." If there was a major error or oversight in ChatGPT's output, you should be able to find it. Wanna try?

 

[bilby]

Google Search, ChatGPT and Wikipedia were my best three choices to get an overview.

Then you had no chance of getting a worthwhile overview, and should have accepted your inability to do so.

Yet instead you were tempted, and succumbed.

It remains the indisputible case that bullshit does not improve the value of any discussion, and that ChatGPT only generates bullshit.

If there was a major error or oversight in ChatGPT's output, you should be able to find it.

Why? I am not in any way qualified to find such an error. Nor is ChatGPT. I am, at least in principle, able to do so (unlike ChatGPT), but I, just like you, have neither the time nor the inclination.

And bullshit need not be erroneous. Error is easier to deal with than bullshit, because error is always wrong, while bullshit has no more need to be wrong than it does to be right.

An unreliable source should be eschewed, even when (perhaps particularly when) it opines on a topic on which we are unqualified to assess its accuracy.

Bullshit is unconcerned with its own truth value. Bullshit is just, as a wise man once said:

what is "out there on the 'Net", perhaps errors and all

I would omit "perhaps" from that. And it is the indifference to those errors, and even to the possibility of their existence, that is the issue here.

And it will get far, far worse. Because today's bullshit is already acting as the input to tomorrow's bullshit, in a positive feedback loop of increasingly confidently stated "information" that is increasingly unreliable.

Wanna try?

No. Nor do you. If you did, ChatGPT would have been superfluous.

Indeed, that's my whole point. ChatGPT is always superfluous. Despite the many and varied efforts by its devotees to pretend that it is not.

 

[Politesse]

No. ChatGPT does a fair job -- much better than ignorant Swammi could do operating Searches manually even spending an hour or three -- of summarizing what is "out there on the 'Net", perhaps errors and all.

Asking ChatGPT is not unlike Googling an issue and uncritically taking the first four hits as gospel. On simple questions and answers, it will likely do okay. For complex or controversial issues, it is worse than just asking your neighbor.

 

[Swammerdami]

No. ChatGPT does a fair job -- much better than ignorant Swammi could do operating Searches manually even spending an hour or three -- of summarizing what is "out there on the 'Net", perhaps errors and all.

Asking ChatGPT is not unlike Googling an issue and uncritically taking the first four hits as gospel.

No. It's more like taking the first 12 hits and having them amalgamated automatically, sometimes with silly errors. NOT a good approach for legitimate scholars -- though more than a few do just this these days -- but much better than an ignorant layman like myself could do without some hours of effort.

On simple questions and answers, it will likely do okay. For complex or controversial issues, it is worse than just asking your neighbor.

So you say. Did you read ChatGPT's responses? If so, can you point to a major error or omission in its responses?

One comment intrigued me. I've never thought that a first Gospel could be written 40 years after the Crucifixion; that an earlier proto-Mark must have existed. The following suggests this is an argument in FAVOR of the post-70 date:

Many scholars believe Mark used oral and written sources, including early passion narratives. This would require some passage of time after Jesus’ death (~30 AD) to develop.

Does this make sense?

 

[Politesse]

much better than an ignorant layman like myself could do without some hours of effort.

If you invest the hour, you now know what you read and learned, remember your thought process as you were learning, your growing insight on the biases of various sites and authors, and will remember much more about the content, having picked it up a piece at a time. It it easier and quicker to use automatic readers to do your thinking for you? Of course. But you get what you pay for. Minimal investment for minimal understanding.

 

[Swammerdami]

Asking ChatGPT is not unlike Googling an issue and uncritically taking the first four hits as gospel.

No. It's more like taking the first 12 hits and having them amalgamated automatically, sometimes with silly errors. NOT a good approach for legitimate scholars -- though more than a few do just this these days -- but much better than an ignorant layman like myself could do without some hours of effort.

On simple questions and answers, it will likely do okay. For complex or controversial issues, it is worse than just asking your neighbor.

So you say. Did you read ChatGPT's responses? If so, can you point to a major error or omission in its responses?

It it easier and quicker to use automatic readers to do your thinking for you? Of course. But you get what you pay for. Minimal investment for minimal understanding.

So the answer is No, then.

One comment intrigued me. I've never thought that a first Gospel could be written 40 years after the Crucifixion; that an earlier proto-Mark must have existed. The following suggests this is an argument in FAVOR of the post-70 date:

Many scholars believe Mark used oral and written sources, including early passion narratives. This would require some passage of time after Jesus’ death (~30 AD) to develop.

Does this make sense?

And you couldn't answer this one either.

 

[Politesse]

The truth of my statement about ChatGPT is not affected one way or the next by its performance on this particular issue. But I'm happy to have a conversation about the question of Matthean priority if you like (provided it is with a human and not a language generating algorithm) and indeed have already commented on the issue in this very thread.

 

[Swammerdami]

Let me summarize.

In a very brief session, ChatGPT produced what SEEMED to be a thorough and articulate summary of the major arguments for and against the 70+ dating of Mark. Did it make a major error or overlook an important argument? I dunno. We have posters here who purport to be trained on the topic; can (and WILL) any of them identify a major error or omission in the Bot's summary? No; I didn't think so.

Is there an on-line page that offers as good and concise a summary as the Bot generated? Probably. We have posters here who purport to be trained on the topic; can any of them point to a summary as good as the Bots? No?

What I see instead is criticism spewed without even reading the Bot's output, let alone identifying flaws. I could say more, but will content myself with the following for now:

Chatbots are improving rapidly. From a Stanford report in February:

Jonathan H. Chen, MD, PhD and a team of researchers are exploring whether chatbots, a type of large language model, or LLM, can effectively answer such nuanced questions, and whether physicians supported by chatbots perform better.

The answers, it turns out, are yes and yes. The research team tested how a chatbot performed when faced with a variety of clinical crossroads. A chatbot on its own outperformed doctors who could access only an internet search and medical references, but armed with their own LLM, the doctors, from multiple regions and institutions across the United States, kept up with the chatbots.

I've highlighted the phrase "kept up with" where some would hope to see "improved on."
The rapidly improving AI's will continue to emit occasional stupidities but the frequency of such will decline. Few will call this "AGI" anytime soon. Nevertheless, and whether we like it or not, AI's already outperform on some tasks previously doable only by skilled humans.

 

[Swammerdami]

But I'm happy to have a conversation about the question of Matthean priority if you like and indeed have already commented on the issue in this very thread.

It does seem like an interesting topic! and I'd be happy to see your arguments summarized, or at least cited with post number(s). The Bot DID respond to this query as shown above

Did Matthew's Gospel borrow from Mark's or vice versa?

Since the summary suggested Markan priority, you might treat the response generated via LLM as the falsities to refute. If it's not beneath your dignity.

 

[bilby]

What I see instead is criticism spewed without even reading the Bot's output, let alone identifying flaws.

The flaw is the bot; The bot is the flaw.

An inability to identify specific flaws in ChatGPT output is the problem. It is the reason why such machines are both useless and dangerous, and why they should not be used.

If you are qualified to confirm that the output contains no major errors, then you don't need the bot.

If you are NOT qualified to confirm that the output contains no major errors, then you mustn't use the bot.

There is therefore no situation in which it is reasonable, helpful, or safe to use the bot.

 

[Swammerdami]

[I have replaced the word 'bot' with 'human' throughout bilby's message.]

The flaw is the human; The human is the flaw.

An inability to identify specific flaws in output form a random human is the problem. It is the reason why humans of doubtful credentials are both useless and dangerous, and why they should not be listened to.

If you are qualified to confirm that the output contains no major errors, then you don't need the human opinions.

If you are NOT qualified to confirm that the output contains no major errors, then you mustn't listen to that human.

There is therefore no situation in which it is reasonable, helpful, or safe to listen to a random human.

I have seen -- with my very own eyes -- responses in some of these Jesus-related threads that were far stupider than anything I've seen come out of a bot! Your advice -- to distrust the bots' output -- applies in spades to many remarks issued on message boards like this one by humans who pretend that their opinions on certain matters are informed and/or valuable.

 

[bilby]

[I have replaced the word 'bot' with 'human' throughout bilby's message.]

The flaw is the human; The human is the flaw.

An inability to identify specific flaws in output form a random human is the problem. It is the reason why humans of doubtful credentials are both useless and dangerous, and why they should not be listened to.

If you are qualified to confirm that the output contains no major errors, then you don't need the human opinions.

If you are NOT qualified to confirm that the output contains no major errors, then you mustn't listen to that human.

There is therefore no situation in which it is reasonable, helpful, or safe to listen to a random human.

I have seen -- with my very own eyes -- responses in some of these Jesus-related threads that were far stupider than anything I've seen come out of a bot! Your advice -- to distrust the bots' output -- applies in spades to many remarks issued on message boards like this one by humans who pretend that their opinions on certain matters are informed and/or valuable.

If you are expecting me to disagree with the idea that "There is therefore no situation in which it is reasonable, helpful, or safe to listen to a random human", then you are going to be sadly disappointed.

My entire argument here is that ChatGPT is as unreliable as a random human.

I am impressed that you managed to spot my clearly expressed and not in any way concealed argument, but bemused that you believe that your having done so somehow undermines it.

We should reject any use of "AI" as a source of information, just as we should reject any use of random humans as a source of information.

There are lots of stupid people.

As stupidity is not in short supply, we should eschew mechanising it.

 

[Swammerdami]

My entire argument here is that ChatGPT is as unreliable as a random human.

I wrote poorly if you think I am in agreement with that.

The average human is far FAR stupider and much more unreliable than ChatGPT on the type of query which I posed to the Bot in this thread..

And furthermore this message board contains almost nothing except humans of high or very high intelligence, and Yet I said with confidence that:

I have seen -- with my very own eyes -- responses in some of these Jesus-related threads that were far stupider than anything I've seen come out of a bot! Your advice -- to distrust the bots' output -- applies in spades to many remarks issued on message boards like this one by humans who pretend that their opinions on certain matters are informed and/or valuable.

 

[bilby]

The average human is far FAR stupider and much more unreliable than ChatGPT on the type of query which I posed to the Bot in this thread..

Sure. But their errors and uncertainties are much easier to detect. We have evolved to do it. LLMs are built to sound like educated, intelligent, reliable and confident humans; They are, however NONE of these things.

And furthermore this message board contains almost nothing except humans of high or very high intelligence

And ChatGPT doesn't have any intelligence AT ALL.

 

[Tigers!]

Pity the thread is getting derailed about ChatGPT

 

[Swammerdami]

These days I often get Google AI's opinion "for free" when I use Google Search, but other than that I've almost never used any of these chat-bots. BUT the thread topic here seemed like one where a chat-bot might produce a reasonable summary so I decided to experiment with what is by far the longest session I've ever spent with a chat-bot. The "synoptic problem" is not one I want to study or become expert on but my curiosity veers off in a zillion directions, and the many discussions here about early Christianity have provided me with veer direction number one_zillion_and_1 ! (Is this yet another symptom of some form of autism?)

In #23 I give ChatGPT's response to questions related to thread topic. The response seemed sensible and balanced. But of course I lack the expertise to evaluate it. 98% of what little I think I know about Gospel dating and the priority of Matthew came from this very ChatGPT output, which I think -- but obviously do not know -- MIGHT be reasonably competent and balanced for a brief summary. (The bot's IQ is less than that of your average banana slug, but it IS adept at doing quickly what would take me MANY hours of Googling effort.)

Since we have experts here, I'd hoped that one of you might condescend to critique the ChatGPT output; and inform us about the most obvious errors, omissions and biases in that output, shown in #23 above.

much better than an ignorant layman like myself could do without some hours of effort.

If you invest the hour,

I asked the bot 3 or 4 questions; an hour would be barely enough for me to peruse the top Google hit for each well-chosen query; and such a hit is likely to be one-sided. I suppose there are some pages that would compare both sides of a question, but they might not be easy for me to find and -- even if found -- would probably not compete with ChatGPT for succinctness.

Kudos to you @Politesse if you could start from scratch and derive a summary like ChatGPTs in sixty minutes!

you now know what you read and learned, remember your thought process as you were learning, your growing insight on the biases of various sites and authors, and will remember much more about the content, having picked it up a piece at a time.

Thank you for teaching me how to learn! I will need to write letters to several peer-reviewed journals asking them to retract my papers, with teachings from a guy who didn't even have proper learnings!

Have I offended you somehow?

Since you ask, sometimes you have disrespected and annoyed me. Even in this thread you imply that ChatGPT's output might be as good as yours but would rather lecture me about How to learn than evaluate the bot's briefish responses. Due to my own character defects, I may retain an unpleasant memory and later respond in kind.

Shall we shake hands and start afresh?