Piero Bosio Social Web Site Personale

@mlanger @BenjaminHan I'd say it's a really interesting look at the discourse structure of text that is and isn’t recognized during LLM training. So a look at sentence level vs larger inference ability.

@alice @rozeboosje to be fair, I think that most of those issues can be mitigated by using smaller models that have been trained on ethically obtained data using renewable energy, while the mistakes are intrinsic in the technology and will always be therebut there is also a lot of openwashing in the industry, and I wouldn't trust any claim of the training data from the providers of the models unless they can be independently proved, and honestly my interest in using the thing isn't enough to justify the work of verifying those claims, so I don't actually *know* whether those models actually exist.

@mnl this exact process is simplified by applying data modeling heuristics for relational schema abstraction. db-eng grammatical blathering aside, specifically I'm referring to the 'third normal form'. Since I'm on a tablet and doing my best version of Palm Pilot short-hand swiping for word completions, instead of typing like I'd prefer, here's some links that describes what I'm referencing about...- Data schema normalization, 3NF: https://docs.oracle.com/cd/B10500_01/server.920/a96520/schemas.htm- Or for those who consider Oracle resources undesirable: https://www.ssp.sh/brain/third-normal-form/writing code on paper is hardcore, impressive, reminds me that I don't own a pen, though I do have a laser printer in a box somewhere.

The human brain does not desire knowledge, or safety, or truth - it desires certainty. The job of science is to deconstruct and rebuild the brain’s naive concept of certainty.In that light, #LLMs are no different from fortune telling, or religion. What they tell you may be true or it may not - but truth is not the product. *Certainty* is the product.They make you *feel* enlightened. That they can do this at a scale that fortune tellers could never dream of should be terrifying.

@metin @buherator to paraphrase a friend, retention of critical thinking and problem solving skills in an AI saturated world will probably become a superpower for those who can resist the autocomplete revolution

@G3rtAggiungerei rendendo i bullshitter dei semplici esecutori non pensanti.Si renderanno ancora più sostituibili nel momento in cui saranno disponibili ausili controllabili direttamente dalle LLMs.

@stefan ha ha ha

@jzb it's worse: we have built hundreds of machines peeing in the pool. And their reservoir of pee is infinite.

@MugsysRapSheet @Lazarou GIGOWALOE(Garage In, Garbage Out, Wasting A Lot of Energy)