Products > ChatGPT/AI

a lovely AI rant

(1/2) > >>

madires:
I Will Fucking Piledrive You If You Mention AI Again: https://ludic.mataroa.blog/blog/i-will-fucking-piledrive-you-if-you-mention-ai-again/

janoc:
Brutal! But spot on - and I am saying that as an actual product owner responsible for bringing some AI features (not LLM-related) in our product to market currently.

The amount of grifters - and the clueless idiots jumping on every bandwagon seeing dollar signs every time there is another fad or hype wave - is nuts.

Things like blockchain or solar roads were relatively harmless - it "only" wasted (a lot) of money. However, this AI hype pushing poorly performing, overhyped inscrutable (non-)solutions into every corner of our lives has a very real potential to make people miserable when they get rejected for loans (or jobs) because "computer said so!", when they get profiled and pulled off flights or abused by the authorities because "computer tagged you as a risk". Or when people get hurt because someone didn't do their job and instead "outsourced" it to some ChatGPT copy resulting in some bad law, policy or a decision.

This was obviously the case before as well but if you end up on some no-fly list, you have at least a theoretical chance to get the record corrected. If you end up being tagged by some machine learning model that nobody knows how it works (both because it is "trade secret" of the vendor and also because even the vendor doesn't have a clue why the model reasons the way it does), then good luck!

There are plenty of real world practical applications of AI and machine learning (AI != machine learning and even less LLMs, it is a much wider field) that are being successfully used every day - and most people don't know about them because they are not where the hype is and are so common and "pedestrian" that nobody considers them as something worth of an extra mention - "Here, we do use AI!". Good examples are things like image analysis (think medical, surveillance, xray, QA, weapon guidance, navigation ...), language translation, spelling & grammar checks, various controllers (even the basic state machine or things like fuzzy logic are historically considered to be topic of study for the AI field), expert systems based on decision trees, e-mail spam filters, etc ...

Practical and actually useful deployments of LLMs are extremely rare in comparison, despite the billions in capital being poured into the tech.

golden_labels:
I see article’s popularity skyrockets. I got it from one chatroom over 6 hours ago.

I dislike a few things about that rant. But other than that: could be my own words. Down to stressing that I’m neither against machine learning nor try to downplay progress in that field. Quite opposite: a few overhyped smortnets are both obscuring many true marvels and trivializing the otherwise deep subject.

Things I dislike? The tone brings a picture of a teenager, who just discovered how humanity works and gets angry about that. I may be wrong of course, but if I had that feeling, other readers may get the same impression. Including those, who should read this. There is no benefit in us circlejerking.

The part with a bar chart. The author misses a much bigger issue, one with higher priority. The survey is not measuring, what it’s supposed to represent! Yet another example of amateurs only superficially mimicking what real researchers do, but being convinced this is a serious and well-done thing. Kids playing adults, once again.

Finally: making wrong assumptions about what “beneficial” means to managers in companies. It’s very different from what an engineer would consider beneficial.

Bicurico:
What do you think of this text?

The text is a blunt and passionate critique of the overhype surrounding AI. The author highlights how many AI projects are poorly executed or fraudulent, driven by a culture of grifters and incompetence. They argue that companies should focus on fundamental IT competencies rather than chasing AI trends without understanding or need. This perspective is valuable for its emphasis on practical, grounded approaches in technology over sensationalism.

janoc:

--- Quote from: Bicurico on June 20, 2024, 08:03:32 pm ---What do you think of this text?

The text is a blunt and passionate critique of the overhype surrounding AI. The author highlights how many AI projects are poorly executed or fraudulent, driven by a culture of grifters and incompetence. They argue that companies should focus on fundamental IT competencies rather than chasing AI trends without understanding or need. This perspective is valuable for its emphasis on practical, grounded approaches in technology over sensationalism.

--- End quote ---

Sounds like an autogenerated summary by some bot to me.

Navigation

[0] Message Index

[#] Next page