HN user
The art styling is more or less uniform too.
I haven't seen many AI works that produces a pelican on a bicycle done in a "Ligne Claire" style, for example.
I guess AI's narrows down the output probability space drastically and converge on some agreed upon aesthetics. Works great for computer programs but bad for art.
The funny thing is that this is not a linear relationship.
It is not like ah.. there is my 85K job or 150K job and now I will never get fired or laid off.,
Late stage capitalism will find its way to make it more efficient at 85K level or at 1.5M level. Pick your red/blue pill.
But that's the nature of the beast isnt it? A probabilistic token predictor will all always have some errors from a human perspective, more and more energy, money and resources will always be needed to control and direct towards desired outcomes.
It will burn up the tokens to get through the deterministic gates, more so when n order dependencies are involved in the mix. Enough typewritters and monkeys could get it done too.
what is of really good use is a super factory. The super factory drives the factory which builds the harness of harnesses.
3% of measurable gains is massive in itself isn't it? Taken as a percentage of the total volume of workforce and the world economy?
If we can gain 3% across the board gains on AI based tasks without subsidized expenses, that would be a great win!?
It's hard to keep track of how many times American football or Baseball is used as a back drop of some TV shows or a hollywood blockbuster. How many US pop culture figures have performed at half time shows. How many U.S. heads of states have mentioned about their favorite teams and stadiums etc.,
Agree. "Culture" provides the nurturing "gap fill" while the capital is not exactly flowing and rules are evolving over time. Culture keeps the lights on in lean years and builds camaraderie when the sports is thriving.
Cannot keep count on how many times I have heard of "Babe Ruth" (1895 - 1948) mentioned in a movie/tv show entertainment context, where a grandpa or an elderly figure is fondly reminiscing things to their gen X grandkids!
Answer is simple.
Why isn't U.S. better at Field hockey, Badminton, Cricket, Ping pong etc.?
Ideally a capitalistic system is supposed to produce the best or close to the best in each category to stay competitive then what gives?
Capitalism produces the "best" in categories where there is a high demand for a product. In the U.S., the demand is for high-scoring, high-production-value entertainment. This is closely entwined with "culture".
This is why the U.S. leads in sports that are tailored for television and massive stadium revenue. Sports that are more nuanced, low-scoring, or lack a domestic TV or the US cultural zeigeist simply cannot compete for the attention span of the American consumer, and by extension, the capital that follows that attention.
calc, stat etc from a text book is something they would naturally be good at but I don't think book based computations thats in the training set and its extrapolations is what is at question here.
They are not great at playing chess as well - computational as well as analytic.
Well the same logic can be used to justify bias too. I am surge pricing a person of religion X or race Y because I am increasing the supply for All races / religion in a equitable way!
Do you have a blog or a writeup about this?
What would have been the cost if it was not DIY'd? Is this doable only in a rural/semi-urban settings?
As one commenter put it: "You can fake a star count, but you can't fake a bug fix .. "
The way to beautify the pig is to put lipstick on the pig!
One could take anything like a cell and split it into genes, molecules, atoms, sub-atomic wave functions (with infinite value range) and take time which can be split into another infinite entity say even within a finite interval. How does this analysis account for that?
I could split this object into 10^500 or 10^50^500^5000 etc., with imagination being the limit.
These values Id'd at whatever imaginable resolution are far from practically useful but at a cosmic scale, there is no telling what is a useful value?
So this framework seems to be more limiting because we define a resolution ?
> Incidentally, San Francisco public schools had a combined admission rate of just under 20%, well below the state average. Mission had the highest rate (26.5%) and Balboa the lowest (15.4%). It may or may not be a coincidence that 90% (the most of any SF school) of the applicants from Balboa were Asian whereas only 25% (the fewest of any SF school) of the applicants from Mission were Asian.
In order to promote diversity of the freshman classroom the college needs to suppress merit to achieve their diversity targets?
Sergey's challenge looks like is not in retiring early or with non-work.
We had a high performing co-worker who was scared witless after a lay-off episode and this was not because he was worried about lacking money or loss of prestige., but because he could not come to terms with the simple fact of facing the 9 am on a Monday morning with absolutely no expectations. It freaked so much to not feel the hustle and the adrenaline rush of experiencing the blues Monday morning!?
Another colleague used to drive up to the parking lot of their previous employer, post lay-off., so that he could feel normal., and he did this for well over 6 - 8 months. Pack bags, wave to his wife and family, drive up in his Porsche to the parking lot and I guess feel normal !?
>This was a five-minute lightning talk given over the summer of 2025 to round out a small workshop.
Glad I noticed that footnote.
Article reeks of false equivalences and incorrect transitive dependencies.
I am pretty sure you wouldn't have touched anything from google and meta as well.
Isn't this sort of defense a weak argument by the courts. If your abstraction is to override a well known common usage/function of a term, then the abstraction doesn't hold much water?
What kind of blue collar work gets you $335K? What would be a 10 - 20 year average wage / year look like?
Is it fair to categorize that it is a pyramid like scheme but with a twist at the top where there are a few (more than a one) genuine wins and winners?
Mostly agree but with one big exception. The real issue seems to be that the figuring out part is happening a bit too late. A bit like burn a few hundred billion dollars [0] first ask questions later!?
[0] https://hai.stanford.edu/ai-index/2025-ai-index-report/econo...
The fact that we now have to write cook book about cook books kind of masks the reality that there is something that could be genuinely wrong about this entire paradigm.
Why are even experts unsure about whats the right way to do something or even if its possible to do something at all, for anything non-trivial? Why so much hesitancy, if this is the panacea? If we are so sure then why not use the AI itself to come up with a proven paradigm?
The title says - "Qualcomm to Acquire Arduino—Accelerating Developers’ Access to its Leading Edge Computing and AI"
Didn't have it on my bingo card that running AI on a microcontroller is what people are salivating for!
Not sure if the strategy is to cram AI into every little shoe box out there and keeping fingers crossed for the stock price to trend upwards!?
Isn't this just fiber optics! Exciting turn of the century tech hitting HN boards in 2025
It is pretty clear that the long horizon tasks are difficult for coding agents and that is a fundamental limitation of how probabilistic word generation works either with transformer or any other architecture. The errors propagate and multiply and becomes open ended.
However, the limitation can be masqueraded using layering techniques where output of one agent is fed as an input to another using consensus for verification or other techniques to the nth degree to minimize errors. But this is a bit like the story of a boy with a finger in the dike. Yes, you can spawn as many boys but there is a cost associated that would keep growing and wont narrow down.
It has nothing to do with contexts or window of focus or any other human centric metric. This is what the architecture is supposed to do and it does so perfectly.
Yes, but usually companies exploit loop holes and build exploitative contracts like you need to pay back the cost that we paid for getting you moved, handled and what not etc., and use it as a threat in case the candidate had the thoughts of leaving before the first year...(anything but the h1b fee, since law explicitly prohibits h1b paybacks). Agree it is going to be a state specific variation of the theme.,
100K one time fee will be easily amortized as a pay reduction over a period of 6 years by the H1B abusing companies.
That is equivalent to getting 5 years worth of salary when you work 6, assuming a median suppressed wage of 100K. This does not seem much of a deterrent for any of these involved.
This could actually result in wage suppression for the victim and nothing else in the long run.
Seems to be poorly thoughout?
Would having a Markov chain of Markov chains help in this situation. One chain does this when 2D bitmap patterns are vertical and another one for left to right?