The Guide Mark II says, “Don’t Panic,” but when it comes to the state of Artificial Intelligence, a mild sense of existential dread might be entirely appropriate. You see, it seems we’ve built this whole AI shebang on a foundation somewhat less stable than a Vogon poetry recital.

These Large Language Models (LLMs), with their knack for mimicking human conversation, consume energy with the same reckless abandon as a Vogon poet on a bender. Training these digital behemoths requires a financial outlay that would make a small planet declare bankruptcy, and their insatiable appetite for data has led to some, shall we say, ‘creative appropriation’ from artists and writers on a scale that would make even the most unscrupulous intergalactic trader blush.

But let’s assume, for a moment, that we solve the energy crisis and appease the creative souls whose work has been unceremoniously digitised. The question remains: are these LLMs actually intelligent? Or are they just glorified autocomplete programs with a penchant for plagiarism?

Microsoft’s Copilot, for instance, boasts “thousands of skills” and “infinite possibilities.” Yet, its showcase features involve summarising emails and sprucing up PowerPoint presentations. Useful, perhaps, for those who find intergalactic travel less taxing than composing a decent memo. But revolutionary? Hardly. It’s a bit like inventing the Babel fish to order takeout.

One can’t help but wonder if we’ve been somewhat misled by the term “artificial intelligence.” It conjures images of sentient computers pondering the meaning of life, not churning out marketing copy or suggesting slightly more efficient ways to organise spreadsheets.

Perhaps, like the Babel fish, the true marvel of AI lies in its ability to translate – not languages, but the vast sea of data into something vaguely resembling human comprehension. Or maybe, just maybe, we’re still searching for the ultimate question, while the answer, like 42, remains frustratingly elusive.

In the meantime, as we navigate this brave new world of algorithms and automation, it might be wise to keep a towel handy. You never know when you might need to hitch a ride off this increasingly perplexing planet.

Comparison to Crypto Mining Nonsense:

Both LLMs and crypto mining share a striking similarity: they are incredibly resource-intensive. Just as crypto mining requires vast amounts of electricity to solve complex mathematical problems and validate transactions, training LLMs demands enormous computational power and energy consumption.

Furthermore, both have faced criticism for their environmental impact. Crypto mining has been blamed for contributing to carbon emissions and electronic waste, while LLMs raise concerns about their energy footprint and the sustainability of their development.

Another parallel lies in the questionable ethical practices surrounding both. Crypto mining has been associated with scams, fraud, and illicit activities, while LLMs have come under fire for their reliance on massive datasets often scraped from the internet without proper consent or attribution, raising concerns about copyright infringement and intellectual property theft.

In essence, both LLMs and crypto mining represent technological advancements with potentially transformative applications, but they also come with significant costs and ethical challenges that need to be addressed to ensure their responsible and sustainable development.