Long horizon events like this on Polymarket stabilize around a % odds corresponding to time value of money. You can get 4% buying risk free CDs for that horizon.
this is not true - if you know for certaint that trump will be president through winning an election in 2028 you can make over 20X your money.
at the end of the day people don't actually believe it, which is why trump is valued little. people who aren't willing to bet with their money on things they say so absolutely aren't serious people.
I am not sure that is a useful principle. I tend to keep an umbrella around in the car regardless whether the forecast calls for rain. Do these people similarly avoid stock markets, insurance, and similar products in the risk space?
your scenario here doesn't really make any sense. one you're conflating getting married with staying together. you could get married and then divorced. two, such a bet wouldn't make sense since it could easily be rigged (and indeed this is what happens with sports betting).
again, if someone says they know the winning lottery and they don't play, they're unserious. nearly impossible to rig, high payout - outcome is of interest to layperson.
The idea was that I figured you’d consider that some things are valuable in ways that don’t involve money. Unfortunately I think the conclusion here is that you actually truly believe that there is nothing that cannot be bet if a suitable market could be formed for it. Of course, I should have realized this, because by placing bets in general you are in fact taking a financial position in the concept of betting markets.
If Trump indeed manages to turn the country into a dictatorship I think winning money on polymarket is going to be the last thing you'll be thinking about
Split brain experiments show that a person rationalizes and accommodates their own behavior even when "they" didn't choose to perform an action[1]. I wonder if ML-based implants which extrapolate behavior from CNS signals may actually drive behavior that a person wouldn't intrinsically choose, yet the person accommodates that behavior as coming from their own free will.
> The patients could accurately indicate whether an object was present in the left visual field and pinpoint its location, even when they responded with the right hand or verbally. This despite the fact that their cerebral hemispheres can hardly communicate with each other and do so at perhaps 1 bit per second
1 bit per second and we are passing complex information about location in 3d space?
Yeah, that sounds very unlikely. The full paper dismisses the possibility:
> Another possible explanation to consider is that the current indings were caused by cross-cueing (one hemisphere informing the other hemisphere with behavioural tricks, such as touching the left hand with the right hand). We deem this explanation implausible for four reasons. First, cross-cueing is thought to only allow the transfer of one bit of information (Baynes et al., 1995). Yet, both patients could localize stimuli throughout the entire visual field irrespective of response mode (Experiments 1 and 5), and localizing a stimulus requires more than one bit of information. Second, [...]
I get the impression that the authors of the paper have some kind of woo (nonmaterialist) view of consciousness. But they also mention this possiblity, which seems more plausible to me:
> Finally, a possibility is that we observed the current results because we tested these patients well after their surgical removal of the corpus callosum (Patient DDC and Patient DDV were operated on at ages 19 and 22 years, and were tested 10–16 and 17–23 years after the operation, respectively). This would raise the interesting possibility that the original split brain phenomenon is transient, and
that patients somehow develop mechanisms or even structural connections to re-integrate information across the hemispheres, particularly when operated at early adulthood.
> I get the impression that the authors of the paper have some kind of woo (nonmaterialist) view of consciousness.
Indeed:
"Our findings, however, reveal that although the two hemispheres are completely insulated from each other, the brain as a whole is still able to produce only one conscious agent."
Which is materially impossible, given the premise.
That's a great paper, but I don't think it calls into question anything about post-hoc rationalizations, and it might actually put that idea on more solid ground.
I don't understand this view (although I hear it often enough): isn't it commonly accepted that the brain and its resulting mind is extremy modular, with all components (trying to) play together?
Noteworthy: this is powered by a new ultra-low-latency userspace TB5 driver offering an ibverbs/RDMA interface over thunderbolt 5[1]. Non-inference HPC workloads could see a massive benefit as well.
I figured one of the really useful applications of LLMs would be in processing transcripts of city government public meetings.
First I was doing it for myself, but I figured others might be interested to read these summaries as well. I use a few prompts to semi-automatically create reports of the most discussed topics for various board/commission meetings.
Ad free, done in my spare time. Just figured people might be interested.
This could be useful for e.g. bazel, where I’ve regularly seen deleting the bazel caches take on the order of 10 minutes because of absurdly large and repeated directory trees (caused by things like runfile trees containing the python interpreter).
It’s somewhat true in the sense that banks need to keep some fraction of their deposits in cash as reserves (the “reserve ratio”). So if they lend money it needs to be backed to that extent by their deposits.
Hash tables are (usually) faster to do all sorts of operations than tree based maps, as most operations become a simple function to calculate a tree’s hash followed by a table lookup. Of course, they’re unordered, so if you need to iterate in order, or find all keys < a certain value, or things like that, tree maps can be better for your algorithm.
Also, TreeMap uses a red-black tree to implement the map, which is a basic type of binary tree. Depending on the data you’d like to store, other kinds of tree-based maps can have better performance characteristics. A map based on a Splay Tree[1] speeds up repeated accesses, so it could perform well if you had keys that were cheap to compute an ordering but expensive to compute a hash, and your access pattern has good temporal locality.
