HNHacker News
TopNewBestAskShowJobs

RosanaAnaDana

2,140 karma · joined November 3, 2017

Water; land cover and land use; remote sensing of forests; biogeochemical cycling; disturbance.
submissionscomments
RosanaAnaDana··on Planetary defense test deflected an asteroid but unleashed a boulder swarm
Wouldn't breaking up a planetary defense worthy asteroid into medium to even large boulder size asteroids be considered a win? Whats the limit on asteroid size that can make it through the atmosphere? And even then, intuitively, it would seem like the destructiveness is going to scale with the size of the individual asteroid, so a rain of many smaller asteroids making it to ground would be far lower than one large asteroid..
RosanaAnaDana··on IBM and NASA open-source largest geospatial AI foundation model on Hugging Face
Ugh puke in my mouth, but I agree. HF is too useful to not be nerfed the toxicity that is modern big tech.
RosanaAnaDana··on What happened in this GPT-3 conversation?
This is an interestingly surreal 'error' mode. So surreal its suspicious. It strikes me that there could be keys to how OpenAI is training ChatGPT in failure modes like this.
RosanaAnaDana··on IBM and NASA open-source largest geospatial AI foundation model on Hugging Face
Think of it like a GitHub for sharing datasets and weights.

I'd like 'huggingface' to the face hugger alien. You can just rip off the front and back of models and train them on new data. Although I've been told explicitly by the HF this is not the case, it's my preferred head cannon.

RosanaAnaDana··on IBM and NASA open-source largest geospatial AI foundation model on Hugging Face
This is such a weird headline and dataset. It's not a very large model, esp for geospatial. And the data set is microscopic, not even 1k image tiles.

A typical geospatial UNET would be trained on any where from 10x to 100x this much data.

This is more like a toy dataset I would give an intern to play on. But to be clear, one would need much much much more data do do something interesting on. Likewise, there are a lot of data filtering and data processing considerations that come into play with satellites like clouds, ascension or descenion, averaging to try and get fewer clouds. Satellite and all remote sensing ML is tricky stuff.

RosanaAnaDana··on LK-99 is an online sensation but replication efforts fall short
or just billions upon billions of dollars.
RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
Its the largest *published* geospatial model. There are plenty of models being developed daily with more than 100m parameters.

I think its just that there is a dearth of publishing on the advances in geospatial ml.

RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
So the big issue is can be summed up as one present across many domains of ML, but particularly challenging in remote sensing. ML has seen rapid advances in spaces where foundational models capture the variance of a domain the latent space of a model. By training on very large dataset, the variance becomes encoded in that latent space; lots of training data implies that the latent space contains meaningful variation. This is what powers things like stable diffusion. A model trained on a very large dataset was connected with another dataset, in latent space. The image model wasn't starting from square 0. It had meaningful variation encoded into it.

Enter remotely sensed dataset. They are monumental in scale. So much so that the 804 chips they submitted here is pretty much laughable. Likewise, some parameters of the data have inherent meaning; the pixel dimensions correspond to real world measurements; the bands are specific spectral windows.

Its similar but far more than just working with RGB cell phone camera data.

RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
If the models aren't available in a repo for download do they actually exist?
RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
Not at 10m brah.
RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
Descarte labs folded?

Oh man. I had no idea! These guys were some of my prime competition for years.

The true cost of venture capitol revealed.

RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
Like... no?

InfSar (InSar, SAR, whatever we're calling it these days) isn't a drop in replacement for anything. Its really neither here nor there when it comes to the utility of other dataset. Infsar is amazing, dgmw, but its stands on its own and has its own advantages/ disadvantages.

The ocs point stands. Satellite data is tough because there is a shit ton of atmosphere between you and the target. That issue doesn't go away with infsar and especially not if it isnt coincidentally collected with higher resolution spectral data. I've been in the industry for around 15 years. Things have gotten better, but really, its important to understand the context and limitations of specific platforms. Afaik, there is no panacea.

RosanaAnaDana··on IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face
I regularly use up to 3cm aerial imagery.

There are many very nice commercial products available.

RosanaAnaDana··on Why has Britain become so poor?
UKs austerity dates back to Thatcher.
RosanaAnaDana··on Norwegian shipping company bans electric cars on board classic ferry route
Tell that to the insurance industry.
RosanaAnaDana··on Norwegian shipping company bans electric cars on board classic ferry route
Sometimes. Other times when they aren't confident in their estimate of the risk, they respond extremely conservatively.
RosanaAnaDana··on Norwegian shipping company bans electric cars on board classic ferry route
This seems relevant. So assumedly they would have lithium ion fire fighting equipment on board as well as the appropriate training. At a max capacity of 6 vehicle that hardley seems like significantly more training or equipment. Maybe a couple extra CO2 fire extinguishers.

With these details it seems much Ado about nothing (click bait).

RosanaAnaDana··on Norwegian shipping company bans electric cars on board classic ferry route
itt: Tell me you've no clue how a ferry works without telling me you've no clue how a ferry works.
RosanaAnaDana··on Norwegian shipping company bans electric cars on board classic ferry route
There are special chemical fire extinguishers to deal specifically with this. I'd bet dollars to donuts this came out of an insurance exercise where it wasn't necessarily the scale of the risk, but rather the uncertainty around the risk (because of its novelty). Insurance companies are very cautious about uncertainty, even if the risk seems low.
RosanaAnaDana··on Autoenshittification. How the computer killed capitalism. – by Cory Doctorow
Also, the distinct and clear path it carved that went directly towards enshittification. Its not (only) that its so shitty, its that it once represented a truly better thing.
RosanaAnaDana··on Lab-grown cotton is on its way to consumers
Cotton or linen only in hot weather. Not that thick fruit of the loom stuff but some good fabric that is thin enough to breathe, but it's still cotton so it's wicking moisture. Evaporation results in a cooling effect so long as the cotton is thin enough. That's the only way to go in the heat (imo). Cheap old men's dress shirts from the 80s that are still 100% cotton and have been washed too many times.
RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
My clients are home owners who are looking to do what it takes to avoid a non-renewal and mitigate/ reduce their personal wildfire risk. We give them the reports and data necessary to challenge a non-renewal. He can hop on our website right now and schedule an inspection. We service the entire bay area.

But I can tell you this right now, with what I saw in the video he won't pass. He's storing combustibles too close to the structure. It puts him at significant risk.

At my company we see customer service as the top priority and try to partner with homeowners to both understand their risk and take action. This is why we focus our reporting so much on education around the issue and communication about what matters for risk.

Ultimately however, the onus is on the homeowner to take action. It's no one's responsibility but theirs to understand and mitigate their risk. I think many home owners are naive as to how high their risk really is. Look at the reaction to Oregon's recent risk maps. The maps were accurate, and the public panicked. Now they are wordsmithing the science to down play the risk. Rearranging the deck chairs does nothing about the underlying hazard.

Link to the Oregon story: https://katu.com/news/local/oregon-lawmakers-insurance-commi...

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
The safer from wildfire standard and the wildfire prepared home standard would both identify the storage of those kinds of materials within the 0-5 foot zone as a hazard significant enough to not pass.

To you it may not seem significant but to a wildfire risk specialist, these are precisely the kinds of hazards were looking for.

The standards are high, but they are so for a reason. We're losing thousands of homes per season to wildfire.

Was it bad customer service to just drop him? Yes. Should he have the opportunity to mitigate? Absolutely.

But ignorance of your risk doesn't mean some one else is obligated to take the other side of that bet. I've said it a couple times else where. If insurance companies thought they could still be writing policies in California, they would. It's billions of dollars in annual premium they are passing on. They aren't doing so because, yes, storing tires against your house is significant enough of a risk factor to warrant non-renewal. Companies like State Farm would rather pull back entirely than accept the risk.

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
> But what is “debris”? That could be anything.

Yep. This is something we've been working to address at my startup (https://wuuii.com/). We build software for doing these kinds of inspections. Getting property owners to take action is one of the critical steps in the process, but communicating clearly what needs to be done is something we focus on. We use a combination of aerial imagery and on the ground inspection to identify all of the issues and make specific mitigation recommendations.

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
Manned aircraft still way cheaper at scale. You can do whole counties in a day. Fly it 2-3 times a year and sell it to hundreds/ thousands of customers.

Unless the property is ridiculously expensive, don't expect any one to commission a flight for it.

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
Usually you use something like Nearmaps, Hexagon, or Vexcel.

These companies use multi camera oblique imagery that can get views down the side of houses were you can do a 'good-enough' job assessing at least some of the major factors for wildfire risk. Debris or outside storage would fall into this category.

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
This is nothing new for insurance. Aerial imagery has been in use for decades to address different perils. At least this homeowner had someone take a look before they were denied coverage. Most insurance companies will just write off whole zipcodes as 'lost' due to wildfire risk rather than get into this granular a level of detail.

The issue California faces isn't a technical one. Its a social/ behavior issue. Californians will need to change their behavior with regards to how they manage their personal property or the risk will continue to be too high for insurance companies to bear. These properties aren't struggling to get insurance because suddenly we have aerial imagery; they are struggling because the data regarding wildfire risk has become impeccably clear.

Individual actions can make an impact to their individual and community wildfire risk, but fundamentally wildfire risk is a network phenomena. It has to happen at scale. If a homeowner isn't willing to deal with debris piles on their property they are putting themselves and their community at risk. Should they get insurance?

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
No sane company wants to leave a market the size of California on the table. They are leaving because the risk is just too high. There is nothing Commissioner Lara can do about the risk analysis associated with the previous 20 years of wildfire.

IF Californians don't change their collective behavior around addressing the issues related to wildfire risk, its unlikely they'll be able to maintain coverage.

RosanaAnaDana··on Insurance company drops customer saying a 'drone' took photos of clutter in yard
Disclosure: I am a cofounder of a startup in Northern California focused on wildfire risk inspections and making recommendations to homeowners about what they can do to reduce their risk. I'm on the science side with a focus on the impacts of risk mitigation.

Dealing with outside storage and debris are part of both the Safer From Wildfire standard and the Wildfire Prepared Home standard. Addressing ember accumulation areas (read: any where leaves/ dust/ crap piles up) adjacent to a structure or any where else on a property is one of the easiest things a home owner can do to reduce their overall risk. Its cleaning up your yard and moving things that might catch on fire away from your house. If you have a problem with that and live in even a mild wildfire risk area, good luck maintaining coverage. This is the absolute lowest hanging fruit with regards to reducing wildfire risk.

Sources:

https://www.insurance.ca.gov/0400-news/0100-press-releases/2...

https://ibhs.org/wildfire-prepared-home/

https://wuuii.com/

RosanaAnaDana··on Is shareholder capitalism a suicide pact?
I think this is a total misunderstanding of how value is created in relationships. The value created can be greater than the sum of its parts.
← PreviousPage 2 of 26Next →