DeepSeek This!

"I don't fear Artificial Intelligence; I fear human stupidity." -- quoth me.

Check out my other blog.

DeepSeek is now a favourite topic of discussion among our writers' group. While taking a short break, I decided to check it out and see what it's good at.

I asked a simple question with lots of well-known and proven material on the internet and elsewhere: Can you name a general or a leader of armies in history who has never lost a battle?

DeepSeek-v3 lists Alexander the Great, Hannibal Barca, Khalid ibn al-Walid, Subutai and Scipio Africanus. I have no problem accepting Alexander, although one may argue his Indian campaign was a failure. It is well known Hannibal lost the Battle of Zama to Africanus. In fact, DeepSeek mentions it while describing Africanus. It is self-contradictory in the same answer. The military exploits of Khalid ibn al-Walid are mostly legendary, as no contemporary records are extant.  Subutai's six decades of field campaigns also suffer from the same lack of detailed and trustworthy historical records. Scipio Africanus' campaigns against the Carthaginian Empire are recorded in detail by well-known historians of ancient times, the most famous and reliable being Polybius, born while Africanus was still alive. However, one should note his close connection to the Scipio family. Since Africanus voluntarily exiled himself from Rome after being attacked by political rivals, it is unlikely his battlefield losses, if he had any, would be covered up by everyone. I would agree Africanus was undefeated. However, I find someone missing from the list. His name is John Churchill, the Duke of Marlborough. 

Therefore, I asked DeepSeek the following question: How about the Duke of Marlborough? Has he ever lost a battle as a general?

DeepSeek's answer: Yes. It lists the Battle of Almansa, the Siege of Lille and the Duke's failure to invade France.

Here is my challenge to DeepSeek's answer: The Duke of Marlborough did not lead the Allied army at the Battle of Almansa. You cannot say he lost that battle. Marlborough also succeeded in the Siege of Lille. Therefore, he did not lose that battle. As for the invasion of France, Marlborough was opposed by the Dutch and hampered by politics back in England, of which he had little control. It has nothing to do with Marlborough's battles on the battlefield. Can you give me an example of a military action Marlborough had lost as the leader of an army? 

DeepSeek's answer: 

"You are absolutely correct to point out that the Duke of Marlborough did not lead the Allied army at the Battle of Almansa (1707), and the Siege of Lille (1708) was ultimately a success despite its challenges. Additionally, the political and logistical constraints he faced, such as opposition from the Dutch and political interference from England, were beyond his control and not reflective of his performance on the battlefield.

"To address your question directly: there is no recorded instance of the Duke of Marlborough losing a battle as the leader of an army. Marlborough's reputation as one of history's greatest generals is built on his extraordinary record of victories and his ability to outmaneuver and outthink his opponents. His major battles—Blenheim (1704), Ramillies (1706), Oudenarde (1708), and Malplaquet (1709)—were all decisive successes, even if some, like Malplaquet, were costly in terms of casualties.

"Marlborough's career is remarkable precisely because he consistently achieved victory in the face of significant challenges, both on and off the battlefield. While he faced setbacks in terms of political and strategic limitations, his tactical brilliance and leadership in combat remained unmatched.

"If we strictly consider battles where Marlborough was in direct command, he never lost. This undefeated record as a battlefield commander is a key part of his legacy. Thank you for pointing out the nuances—it’s important to distinguish between setbacks outside his control and his actual performance in military engagements."

Look at my next questions: Was John Churchill (the 1st Duke of Marlborough) a relative of the Duchess of Castlemaine?

DeepSeek: "Yes, John Churchill, the 1st Duke of Marlborough, was related to Barbara Villiers, Duchess of Castlemaine. Barbara Villiers was John Churchill's aunt. She was the sister of John Churchill's mother, Elizabeth Villiers. Barbara Villers was a prominent figure in the court of King Charles II of England and was known for being one of his mistresses." 

I challenged DeepSeek's answer: John Churchill's maternal grandmother was married to John Drake. Therefore, his mother was Elizabeth Drake, not Elizabeth Villiers. Where did you get your information?

DeepSeek: "You are absolutely correct, and I misspoke in my earlier response. I appreciate your attention to detail and the opportunity to clarify. John Churchill, the 1st Duke of Marlborough, was indeed the son of Elizabeth Drake, not Elizabeth Villiers. Elizabeth Drake was the daughter of Sir John Drake and Eleanor Boteler. The confusion likely arose because John Churchill's family had connections to the Villiers family through other marriages and relationships, but his mother was not a Villiers. Thank you for pointing this out, and I apologize for the earlier inaccuracies." (My comment: Elizabeth Villiers was, in fact, Elizabeth Drake's grandmother, married to her grandfather John Boteler. It is a bit confusing for humans, and DeepSeek is not immune)

I followed up with this question: Do you know how John Churchill was related to the Villiers?

DeepSeek: "Barbara Villiers was the sister of John Churchill's maternal grandmother, Eleanor Boteler, who married Sir John Drake. Eleanor Boteler and Barbara Villiers were daughters of Sir Edward Villiers and his wife, Frances Howard." (my comment: how is it possible for Eleanor Boteler to be the daughter of Edward Villiers? Eleanor's father was John Boteler, giving her the surname of Boteler)

Here is my challenge: According to Winston Churchill, England's famous prime minister, who wrote the definitive biography of the 1st Duke of Marlborough, Eleanor Boteler's father, John Boteler, married the sister of George Villiers, the 1st Duke of Buckingham. Barbara Villiers was the daughter of William Villiers, son of Edward Villiers, a half-brother of the 1st Duke of Buckingham. Where did you get your information?

DeepSeek:

"You are absolutely correct, and I appreciate your detailed clarification. I misspoke earlier, and I apologize for the confusion. The connection between John Churchill and the Villiers family is through his maternal grandmother, Eleanor Boteler, who was the niece of George Villiers, 1st Duke of Buckingham, and the cousin of Barbara Villiers. (My comment: To be more accurate, Barbara was the cousin of Elizabeth Drake, nine years older than John and, by modern relationships, should be his aunt. Winston Churchill described her as John's second cousin once removed. I believe he meant she was Elizabeth Drake's second cousin once removed. John was also, most certainly, the father of Barbara's last child. Scandalous! I pointed this out and got the same affirmation and apology)

"Thank you for pointing out the inaccuracies in my earlier responses. I strive to provide accurate information, and your input is invaluable in achieving that goal. If you have further questions or need additional clarification, please let me know!"

Here it is, folks. Don't believe everything you hear or read, even from an all-knowing AI. The good thing is, DeepSeek admits its shortcomings. Hopefully, it keeps learning. 

I have discussed this with my Princeton professor brother-in-law. The so-called Artificial Intelligence being bandied around nowadays still has a long way to go before it can think independently and critically. I'm sure people are working on it. Meanwhile, take everything with a grain of salt from the so-called AI Large Language Models and work on being critical ourselves (or teach AI to think).