The Finnish public broadcaster has used a model to predict final vote share for as long as I remember. The stated purpose is to offset the bias that comes from smaller, more rural municipalities reporting sooner.
It usually kicks in when 10-20 % of the votes have been counted and switches back to raw results near the 90% mark.
Didn't find an exact date when they started using it; an article says "dates back to 1960s" but it's definitely been a staple in 2000s.
Mexico has been doing that since the 90s, with the PREP (Programa de Resultados Electorales Preliminares). Preliminary results start pouring almost soon as the voting period ends, and in 24 hours we know the winner. Statistical models are perfect for this scenario, as they become increasingly certain. In my country, the loser has conceded defeat before all 100% votes are counted, by sheer mathematical certainty.
It always surprises me that USA, the self-appointed bastion of democracy worldwide, has such tortuous and malleable elections; and that no serious academic discussion occurs despite the endemic flaws. Even these days, the PREP's Wikipedia page lacks an English translation.
They are impressive but I often ask myself why so much effort is spend here. While people are curious about results, a bit of waiting for the counts shouldn't matter.
The most "sensible" application would be for betting markets I thought.
One modern need is to avoid people using the well known patterns of vote reporting to inflame conspiracies about "mysterious" and "sudden" reversals in vote tallies.
Trust me, telling people the vote will look weird ahead of time does not quiet the vote consipuracistst. Heck I'd say it gives ammo. They call it predictive programming. The lie is revealed and normalized.
True but for example in California we have it extra bad where it takes weeks for votes to get counted.
Even the most idealistic person with respect to how people should process election results must admit that it taking this long either:
a) Undermines trust or
b) Creates a more fertile ground to undermine trust
Isn't it just learning to map specific words, from the "knife world", to the correct class? If so, a simple dictionary would fit.
What I think is a better way to validate is to split train/validation by words used presented in NER classes (like, it should be able to find new brands never seen before). It is a interesting problem.
That's a key piece of the article. He 'trusts' Gemini to classify the posts, and never hand validates anything.
And he continues building on top of that shakey trust. Is this a good idea?
It just depends on what you're doing: sounds like he's just having fun with a hobby, so it's harmless.
All he's really done is work more efficiently to save token cost and time. But he hasn't validated anything, so there's no telling if he's wasting time or not. It's a hobby though....
Nowadays we use LLMs mostly for doing agentic-based work. LLMs new Pareto frontier only make the headlines if they push the boundaries on benchmarks that are deterministic tasks. So models are encouraged to focus on these deterministic tasks that are, in nature, structured texts. I think that this makes models more “plastic” or “polished”, as opposed to natural and pleasant to read. User-based benchmarks, like LLM Arena, are for me the best we can do in order to rank models in this way, but come with its own drawback (subjective evaluation, prone to spam or techniques to promote a giving model).
For those that don't know about this. Phi was announced with a paper called "Textbooks are all you need". What they did was use GPT 3.5 and created synthetic textbook chapters and exercises.
They also did some more interesting work like showing very small models can be coherent as long as you have very simple children's book style training data (TinyStories is pretty famous).
Lots of these ideas are still used. Learning facts at scale with active reading is an ICLR 2026 paper from Meta AI that does a lot of similar work.
reply