Hacker Newsnew | past | comments | ask | show | jobs | submit | lp251's commentslogin

why did they move to fblearner

what is the new training platform

I must know


Meta has been itching to kill FBlearner for a while. Its basically an airflow style interface (much better to use as a dev, not sure about admin, I think it might even pre-date airflow)

They are mostly moved to MAST for GPU stuff now I dpn;t think any GPUs are assigned to fblearner anymore. This is a shame because it feels a bit less integrated into python and feels a bit more like "run your exe on n machines" however, it has a more reliable mechanism for doing multi-GPU things, which is key for doing any kind of research at speed.

My old team are not in the super intelligence org, so I don't have much details on the new training system, but there was lots of noise about "just using vercel" which is great apart from all of the steps and hoops you need to go through before you can train on any kind of non-opensource data. (FAIR had/has thier own cluster on AWS, but that meant that they couldn't use it to train on data we collected internally for research (ie paid studies and data from employees that were bribed with swag)

I've not caught up with the drama for the other choices. Either way, its kinda funny to watch "not invented here syndrome" smashing in to "also not invented here syndrome"


Los Alamos is doing well. That’s about it.


Los Alamos has a single high school in their district, and it isn’t even close to the top ranked school in the state.

As I said, the variability is the key metric.


It helps to have a super high percentage of nuclear physics PhD in town :-)


This is paid for by the oil and gas boom in Southern NM.

The medical situation is getting worse by the year, though. I don’t think it’s just a matter of shoveling more dollars


I went on a road trip through Southern NM a couple years ago. Highly recommend stopping at Gila National Forest - it's a certified "dark park", remote enough from sources of light to see the Milky Way with the naked eye.

One thing that struck me - towns down there had a template. 90% of towns we drove through were just a blood plasma "donation" center, a dollar store, a gas station, and a cemetery. Very bleak existence out there, oil and gas boom notwithstanding.


That is a very common pattern. Probably 90% of poor rural America is like that. And the lucky towns have those stores, others not even.


That sounds amazing. It is now on the travel list!


What I was told, was it was anti-undocumented stuff.

People don’t want immigrants getting help that residents pay for, so they turn the spigots off for everyone.


I haven't seen anything like that. The biggest factor you can clearly point to is that NM has some of the lowest salaries for doctors combined with some of the highest medical insurance premiums.


If you live there, and don't agree, then I stand corrected. I was told this by someone that lives there, but that's a sample size of 1 (from Albuquerque).


The icing on the cake there was that they just gerrymandered NM recently in a way that took all representation away from southern NM.

I know other states have as well, so nothing new there, but seeing as they basically fund all the state's social projects, felt a bit done wrong.


shame they won't be as good


People won't like how you said it but there is truth to it. Pretty much all students from my uni who went on exchanges to the US said the level over there is much lower and they were way above the local students.

This was not the case elsewhere, most notably in Asian countries.

There is some truth to every stereotype.


Given how easy it is for rich people to buy grades for their children in most countries, we might see an improvement.


wonder if they still train all of their models using Mathematica because it was impossible to get pytorch on the classified systems


AFAIK that was mostly due to a silly detail about MD5 hashing being restricted on FIPS compliant systems? Or something like that. I'm pretty sure there's an easy workaround(s).


there were a bunch of reasons. couldn’t bring compiled binaries onto the red, so you had to bring the source + all deps onto a machine with no external internet.

it was unpleasant.


Just have Hegseth run PyTorch for them


>pip install *


it does


off topic, but how is living in Estes Park?


ye, that’s a thing. don’t pluck your nose hairs.


I moved from Santa Fe County (Rancho Viejo area) to the Denver metro area and am sad every time I go outside at night. We could see the Milky Way most nights.

The NM sky is amazing during the day, too. Such a vibrant blue!


Did you ever fully acclimate? After 3 years at 6000 ft I still had lower blood oxygen than at sea level. (per apple watch, so take that with a grain of salt)


I can only speak anecdotally about my experience when I was young and now that I am much older. When I was young I went from about 69' MSL to about 6780' MSL when doing a job at the Air Force Academy and we dared each other to run up some stairs. We made it about half way up and fell over laughing and gasping for air. Our team chief that smoked about a half a carton of cigarettes a day walked right past us shaking his head. For about a week or so I had some pressure in my leg near the hip but it cleared up.

I am much older now and a couple years ago I retired/moved from about 600' MSL to 6300' MSL and it took about a half of a year for my body to adjust including some pain in my legs with some associated popping and other weird pressure like feeling similar to what I had when younger but it took longer to go away. Probably too many decades behind a computer and not enough walking. I've had no feelings of self defenestration.


Yea, we would only need a small percentage of folks unable to acclimate to create a significant correlation.


Isn’t Apple Watch and other blood oxygenation just measuring saturation (a relative measure)?

Among other things, Acclimatization involves producing more blood-oxygen carrying capacity (or less if you go back to sea level), which is an absolute measure, not a relative one that the oximeters would pick up.


I haven't measured my blood oxygen, alas.

I've been here 4+ years now, and other than being 4 years older and just coming out of a sedentary 3 year period, I notice zero effects of living (and running, cycling, swimming, hiking) at altitude. Can't say any more than that really.


Hmmm, given that...it does seem to me the initial claim is likely to be true.

I know you said that it has to vary between multiple individuals at altitude, but it's unclear to me why.

Don't we just have to show some individuals don't ever get back up to the same blood oxygen levels?


Well, another comment here from a guy of Sammi ancestry notes potential polymorphism in GHC1 as possible (and likely) partial explanation.

I would imagine that there are a variety of potential genetic differences that could generate a wide range of altitude-adaptation differences.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: