ChatGPT Goes Nuts And No One Knows Why
On Tuesday this week, OpenAI’s ChatGPT was speaking its own language when responding to user inquiries, and no one (except OpenAI programmers) knows why. Though the issue was quickly addressed and corrected by OpenAI, why are all of the social media posts regarding the issue suddenly unfindable?
ChatGPT Starts Speaking A New Language
Before the X posts were deleted Sean McGuire, a senior associate for Gensler posted some screenshots showing just how wild ChatGPT had gotten with its responses.
One shot read, “Sometimes, in the creative process of keeping the intertwined Spanglish vibrant, the cogs en la tecla might get a bit whimsical. Muchas gracias for your understanding, y I’ll ensure we’re being as crystal clear como l’eau from now on.”
What Was The Reason?
The response got a little more weird when ChatGPT said this, “Would it glad your clicklies to grape-turn-tooth over a mind-ocean jello type?” Hmm? Well … would you?
ChatGPT then followed up with a few Bill Evans (jazz pianist) references before getting stuck like a broken record saying over and over again, “Happy listening!” Sounds like Johnny 5 (Short Circuit 2) is running low on battery acid, if you catch the drift.
One X post did make the cut, however, and it was a poll posted by Gary Marcus asking people what they think happened to make ChatGPT go so wonky on Tuesday.
Was There A Hack?
Did OpenAI’s ChatGPT operation get hacked? Did the program have corrupted weights? Was it a hardware issue that caused the bot to be a jumbled mess for a brief time, or was it something else?
Most answers to the poll went towards corrupted weights being the issue, but no one knows with any certainty.
Why The Secrecy?
There are plenty of reasons to speculate why OpenAI is so closed-mouth about the internal workings of their chatbot creation and why a great number of social media posts disappeared so quickly.
If everyone already knows about the boggle, then why work so hard to erase talks of it from social channels?
Just An Easy Answer?
One very compelling reason for secrecy is that OpenAI has (for the past couple of years) been held in high regard as a leader in the AI revolution.
Billions of dollars have been invested in the company since the launch of ChatGPT, and it’s embarrassing for a snafu to be so publicly displayed. The U.S. Military even employs the technology of OpenAI in various ways.
The real story behind what made ChatGPT stray from reality a bit was likely quite simple, as most tech tumbles are.
Need More Transparency
But the general public isn’t as forgiving as a person who spent the last 17 hours trying to program a new piece of software to do exactly what they want it to do exactly when they want it to do it.
The issue certainly would not have been as perplexing to users if OpenAI was a little more transparent about how ChatGPT works. Understandably, they want to protect their “creation”, but everyday users need to know a little more about what they’re working with when they choose to utilize ChatGPT’s abilities.
