thewopr

I'll also add, most food products down there are "expired" already. When I was down there, it was often a challenge to find the oldest piece of food. I think we found 5 years + beyond the Best By date.

Best by dates for shelf stable/frozen food are often not safety related, so the antarctic program just charges forward with whatever they have.

(creds: I was down to McMurdo as a researcher three times)

I suspect this has to do with space and weight constraints, and probably a touch of old-school procurement practices.

In the not-too-distant past, basically everything was flown to south pole station, so weight was at a premium. Powdered milk weights a lot less than UHT milk. Now they do a traverse to the pole with sleds and tractors, so weight is less of an issue, but volume might still be.

On top of that, procurement may be slow to change. If, in fact, weight is no longer a constraint, it might take years for procurement to change to include buying UHT milk.

My experiences were with the McMurdo Dry Valleys LTER program [1]. They have (things may have changed) a number of sites sending telemetry back to the states via the Iridium network. Nothing too fancy. It worked. Biggest challenge really was that iridium was slow and relatively high power (and somewhat flaky deep in the valley where we were). I also had some involvement with the NTL LTER program[4], but that type of work has even easier telemetry constraints (these days, just use a cell modem).

I totally agree with you on the "using a combination of different sensors and a cleverly trained algorithm to get at the parameters of interest". This is something not too far from, in a way, how many sensors work already. They are *proxies* of the actual thing being measured. From my world, the s-can DOC sensor was always a good example, using in-situ spectroscopy to estimate DOC concentration.

Crux of the challenge is "what is the parameter of interest" and "can you come up with a way to estimate it with something easily measured?

Because this is HN, I'll say there is another interesting route possible. If you can change the economics of a situation and decrease the cost of a basic sensor, then you can often increase the volume of applicable uses. I was tangentially involved with the development of the miniDOT [3], which ended up being one of the first "inexpensive" (as in less than $5k) dissolved oxygen sensor. It really changed how people used them and increased the amount of DO sensing by probably an order of magnitude.

[1]: https://mcm.lternet.edu/ [2]: https://www.s-can.at/en/product/carbolyser-v3/ [3]: https://www.pme.com/new-products/minidot-usb-oxygen-logger [4]: https://lter.limnology.wisc.edu/

Having been in technology, antarctic research, and remote monitoring of environmental systems down there, my take: don't read into this this too much.

This is mostly a PR piece, probably pushed by the university or non-profit researchers involved. They are trying to use some sort of partnership with NVIDIA (as loose as that partnership might be) to draw attention and show they are having "Broader Impacts" for their impact statement.

Most eco research is based on historical comparisons of months/years/decades of data So the use of real-time/streaming data down there is pretty limited. You can just as easily shove the data into storage and have a researcher pick it up next time they go down (often *much* easier as you don't have to worry about powering comms systems).

Climate/weather data may be different, if only because some of the data might go into current/real-time weather models. But even there, it's probably a stretch (I know of very little work being done with anything near real-time as far as data goes down there).

I'd argue, they aren't doing something future-proof right now because the fundamental architecture of the LLM makes it nearly impossible to guarantee the model will correctly respond event to special [system] tokens.

In your SQL example, the interpreter can deterministically distinguish between "instruct" and "data" (assuming proper escape obviously). In the LLM sense, you can only train the model to pick up on special characters. Even if [system] is a special token, the only reason the model cares about that special token is because it has been statistically trained to care, not designed to care.

You can't (??) make the LLM treat a token deterministically, at least not in my understanding of the current architectures. So there may always be an avenue for attack if you consume untrusted content into the LLM context. (At least without some aggressive model architecture changes).

This is interesting. Pondering about this, the vulnerability seems rooted in the very nature of the LLMs and how they work. They conflate instruct and data in a messy way.

My first thought here was to somehow separate instruct and data in how the models are trained. But in many ways, there is no (??) way to do that in the current model construct. If I say "Write a poem about walking through the forest", everything, including the data part of the prompt "walking through the forest" is instruct.

So you couldn't create a safe model which only takes instruct from the model owner, and can otherwise take in arbitrary information from untrusted sources.

Ultimately, this may push AI applications towards information and retrieval-focused task, and not any sort of meaningful action.

For example, I can't create a AI bot that could send a customer monetary refunds as it could be gamed in any number of ways. But I can create an AI bot to answer questions about products and store policy.

Is there an English version of the website? I just see the spanish version when I go to https://www.neuraan.com/

Also, as a former DSist in a service area of ecommerce, one challenge with automated service interactions are not just conversations (with answers) but also actions based on those conversations. Of course, something to iterate into, but it may be a question that comes up when talking to potential clients.

Sorry, been a lurker too long. Added my email address to my profile. Also shot you an email to connect.

Actually, that's not my profile. My username doesn't align on github. My github profile is

github.com/lawinslow

As I mentioned, long-time lurker. Added my email address in my profile.

I'm going to finally try to start my own company. Will it be a "startup"? I don't know. But it will be technical and it will go from 0-1 (or from 0-0).

After years of lurking here, watching on the sidelines, working for larger companies, having kids, buying a house, I'm finally going to take the dive. I'm excited, nervous, lost, all at the same time. But I have enough savings and an accommodating spouse, so I don't have to work for a while.

I'm a long-time academic, turned ML-practitioner. I have no major online presence. I don't have a brand. But if anyone is interested in talking, DM me, I have lots of time and am still in the divergent phase of entrepreneurship.

Edit: Added email address to profile. Excuse the confusion, I have been a lurker too long.

It was public knowledge. Further, there is no such thing as New Zealand territory (or anyone's territory for that matter) in Antarctica. There are existing territorial claims, but they are overlapping and basically nullified by the Antarctic Treaty.

https://www.antarctica.gov.au/about-antarctica/law-and-treat...

Hmmm, interesting. Any citation for that number? I've heard numbers like that thrown around, but I don't know where they come from.

This is super interesting and something I've never thought about.

One other issue I have noticed with 401k's. Do 401k's exacerbate inter-generational income inequality? I know a number of people who are going to, in the next few decades, inherit very sizable 401k accounts. This is great, in that their parents were very frugal, saved well, and had comfortable retirements. But on the flip side, a pension would have died with that person, now there is this ongoing inter-generational transfer of wealth that otherwise wouldn't have occurred.

I am a huge Tesla fan (and Elon fan). I own stock (shortly after IPO). I have a model 3 reservation. I have been singing their praises for a long time. Elon is an incredible visionary and probably an incredible engineer. This is a great combination for making the impossible possible.

And I totally agree with this article. If not a new CEO, Tesla needs a good COO. They need excellent, consistent execution, not novel, groundbreaking execution. They have 100's of thousands of reservations for the 3 (and I don't know how many powerwall and solar roof reservations). If they can just execute on this, the world is theirs. But if they continue to have delays and major, public mistakes like the model 3 ramp, my stock purchase may have been a poor choice.

I really wish there was much better details on costs on such large public projects (and institutions). Between the extreme costs of public works like these, healthcare, college education, there is a cost disease that needs to be figured out and addressed.

> More granularity is needed.

I feel like one big step forward on that would be requiring the disclosure of detailed budgets and expenses for all such semi-regulated institutions/projects receiving large portions of their budgets from public funding.

This type of information is already largely relegated to the back of the library. I don't know a mechanic that has re-built a modern engine at this level. Reliability is so high and failure modes are probably generally more catastrophic (someone didn't change the oil) and warrant an entire engine replacement.

Plus, I'm curious if the economics work out at all anymore. Mechanic billed at $100/hour. How many hours does it take to rebuild anything to the piston level, versus the cost of a new (or used) replacement engine just dropped in place. I suspect they favor a replacement engine.

This is the kind of stuff that never makes into a journal article, but frequently gets cooked up by the university public relations/outreach staff. There is a huge incentive to make every piece of work as flashy as possible.

I'll also point out the 10M times faster thing. How was that estimated? If I have some really bad code that I'm using and I optimize it and it runs 10M times faster (not uncommon for scientific code) do I get to make this a big public headline? (the answer is: apparently yes).

In general, it seems like most larger companies opposed this move.

This is a super misleading visualization. I caught wind of this yesterday and started doing some digging. There are some broad issues with this.

1. As others have mentioned, this is them playing with not true misspellings, but the phrase "how to spell X". These are very different things.

2. I don't believe these are actually the true "top searches" in each state. They did some magic here to, I suspect, normalize for the most common of these searches in the English language and then pick the most anomalous search. Otherwise this viz would probably be very boring and just be filled with some of the harder to spell English words.

3. Further evidence this is based on "corrected rank" or something. Looking at the state of Texas, the search for "how to spell beautiful" is much more common than the search for "how to spell maintenance", which their viz claims is the "top" search for Texas. Search for "Beautiful" makes much more sense considering its frequency of use in the English language.

https://trends.google.com/trends/explore?geo=US-TX&q=how%20t...

4. The original version that came out misspelled the word "ninety", which was pretty humorous considering this will be used by thousands to talk about how terrible it is that people can't spell and how autocorrect is ruining america, etc etc.

Of course we should all be shocked that a tweet was somehow unable to convey nuance and detail of a quantitative analysis. /sarcasm

edit: The original source was a tweet: https://twitter.com/GoogleTrends/status/869624196921303040/p...

This is a really interesting comparison. Probably a better comparison than Microsoft or Google, as is done in the article.

Or almost now

$969 for premium specs.

http://www.apple.com/shop/buy-iphone/iphone-7/5.5-inch-displ...

Yeah, that's actually kinda my point. A sequel is kinda the same as additional seasons (or episodes) of a TV show. With a TV show, additional seasons are basically assumed unless it doesn't do well with the target market. But movies are sometimes prepared for sequels, and sometimes not. But apparently, sequels (or additional seasons of a show) are the best way to make money. So movie studios have an inherent handicap in that they tend to create episodes (movies) that have to stand alone, whereas TV, if good, will just gain viewership as they create more seasons.

I'm curious too with your number. How often does it take a series until the second or third season to gain notoriety and a large following. I feel like "Breaking Bad" took into the second season to really get a following. There must be other examples.

Maybe. But the author of the article at least does make it fairly clear what he's arguing for. Not that his diet is the perfect one, but that we need far more research into the hunger drive, not the chemistry of food and sustenance.

While many scientists would scoff at him making this argument outside of the scholarly journal orbit, it's actually a pretty good place to do it. Many scientists I know read broadly, and many decision makers at funding organizations are looking for ideas and perspectives wherever they find them, not just in journals. Especially as, in most journals, this would have gotten promptly rejected as the "study" wasn't scientific. But the point is made and shouldn't be ignored (IMHO).

But what's interesting is that TV is having such a different trajectory (TV as in video packaged as episodes, the lines have blurred). Netflix and the like.

I often wonder if TV is winning because it has built-in sequels. If a series doesn't stick, whatever, money lost. If a new series does well, you can milk it for many many more seasons. TV vs Movies, the risk in cost is the same but the upside is much higher for TV.

Hmmm, I didn't know that. Will have to read. Thanks.

I am not an insider but wow, this really resonated with me.

> We pursued a potentially great summer movie like Edge of Tomorrow and completely botched its release.

Edge of Tomorrow was a pretty solid movie. I would absolutely have gone to it in the theater, but the marketing just did not connect with me. I've run into this with a number of movies now.

I generally agree. I've seen a few challenges to this perspective. 1) Some agencies do not think using 3rd party services like github is allowed. 2) Different groups make the initial code release pretty time consuming (as, in part, it should be) as, through time, the history may have accumulated secrets/keys/other sensitive material. So very thorough reviews are sometimes required. 3) Some agencies allow the use of github (and the like), but make the process of getting code up there extremely challenging. For example, some groups require all changes to code to go through the same internal review process as traditional communication, like a scientific publication or press release (internal review, multiple levels of management, etc).

Of course, the last one is the worst. A lot of groups are getting their act together when it comes to starting open-source, but it is met with a lot of skepticism.

Yeah, absolutely. We make that argument often. In fact, the software we open-source is technically supposed to be released as public domain (as in, no license applied) and often is if people are doing things right.

But that doesn't mean it has to be hosted and released online. Many agencies work under the "not going to release anything until we get a FOIA request" model. Just because it is public domain doesn't mean it is public.

Full disclosure, I'm in a federal department that has been pushing for more open source for a while.

This is a great move by the White House. While there are a lot of groups that are trying to push for more openness and release of software, it can often be challenging. A lot of federal groups have been taught over the years to be very risk averse, and open software is viewed by them to be risk. Probably one of the most common concerns is, "What happens if someone takes and misuses our software?" In a highly risk-averse federal environment, these can be challenging arguments to fight against.

If you like and support this kind of thing, one big thing you can do is to contribute and supply feedback. We frequently have to go to our superiors and justify what we are doing with regards to open source. We say things like, "this repository had X pull requests from non-federal contributors". Or, "We got Y comments and questions from non-federal users of our projects".

It could be as simple as an email saying "Hey thanks, I found this useful", to a full-on pull request fixing an issue or with a new feature request. The more fodder we have to say "open source increases engagement and creates positive feedback" the more you will see this kind of thing happening.