idealmedtech

I'd be very curious to see what the glucose/lactic acid curve looked like during and after the transplant! Usually it's a good leading indicator of broader issues. We've done a few studies in swine (they're the gold standard of blood sugar modelling) and if an animal was not doing well, we'd always see a spike in lactic acid right before death.

When I find incredible in depth content that I plan to read later, my first step is always to drop it into archive.org, and then to bookmark the original link. That way if you're perusing later on and some links have died or been moved, you can always recover the content you want. Plus, it's archived for everyone else on the Internet which is a net benefit.

I donate to them yearly for being such a crucial piece of Internet infrastructure (in the historical, not literal, sense)

That makes sense, thanks for clearing it up!

In the download script, it skips parts of the model (02 and 83); any ML people have ideas why you'd do that?

I think people actually want to read longer form content on Twitter, just look at how often threads appear under the viral tweets topic! It's more fundamentally a twitter problem than a user attention span problem in my opinion.

This is very useful for doing hardware demos to prospective investors/clients! One pane is code and/or UI, other one is live output from your webcam.

I think the important bit here that maybe I didn't make clear is that, at least in the context of men flirting with women, sometimes women are kind and polite without actually being comfortable (social conditioning), which many men incorrectly take as a sign of interest.

I agree that if you're charming and fun, flirting with strangers is pretty straightforward, and I especially agree with the point about getting to know people over a longer period of time before trying to gauge potential interest. Worst case, you get a cute friend!

That's definitely true! It's so much easier to date in person from a logistical standpoint, and I think your comment also hints at what I think is kind of a paradox: in today's social spaces, being a kind and conscientious individual often means giving strangers their space, especially in places that used to be centers of socialization; gyms, coffee shops, parks etc. But by doing this, your opportunities for organic relationships dry up significantly, especially if you're not consistent about being in those spaces. So often times people feel forced into online dating as one of the few "sanctioned" spaces where it's still considered okay to flirt with strangers.

I won't claim to speak for an entire community, but among my many queer friends this seems to be a common thru line! Of course like anything social there's never one clear explanation to any phenomenon.

I see this explanation of the grind of online dating a lot, and I think it misses a more fundamental point: people fundamentally are pretty terrible at talking to strangers. Without conversational skills, a vast majority of online dating conversations simply fizzle out; often times simply due to disinterest, but I think almost always due to a lack of ability to hold a conversation and be both interesting and interested.

I find there's actually a lot of success to be had online dating, if you just know how to talk to people and learn about them! It seems like such a tautology but I think it really is that simple.

And here we see that even this discussion is political; a summary rejection of the premise of the article! At the end of the day, we're all people, and a lot of what we are boils down to politics. You can't exist outside of it.

That's a really neat strategy for ownership! Did you write the contract that you use? Any interest in sharing it?

Previous discussion (694pts, 550 comments)

https://news.ycombinator.com/item?id=31738029

Asynchronous communication is easier for busy people, typically

Would time of day (air temperature) also have effects on it?

It seems like the test suites for these deep learning based models are themselves almost comprehensive knowledge bases from which you could build a more traditional control software.

One is a fundamental human experience. The other is a SaaS.

I find it's best to write readable, straightforward code at first, and when performance really matters, benchmark to find places where you can eke out the percentage points that matter. Premature optimization can waste valuable hours when you don't know what how production workloads will stress your application.

Forgive my ignorance, web et al is not my specialty, but how does this work with HTTPS traffic? Do you install a root cert in the dev environment like Wireshark to be able to mirror agnostically? Or does there need to be application config (DNS change, internal routing changes) to make this work? If the latter, what's the value add here?

There's a huge variety of experiences to be had while woofing, so make sure you prepare for the worst (tent sleeping, very little shower access, controlling hosts). Definitely talk to alums of whichever site you have in mind if possible!

Are we coming full circle back to Knowledge-Based Systems of the 1980s, with a sprinkling of NN?

LEO is too susceptible to interference (of the physical, not radio, kind) for any serious digital communications infrastructure to be placed there. I could see laser-based swarms of comms satellites being used (cubesat size, not starlink size), but I don't think the economics are there quite yet.

I'm a cofounder over at Ideal Medical Technologies, and we're working towards building one of the first fully automated blood glucose control devices for hospital use, which could save hundreds of thousands of lives per year in the US alone!

We're not currently hiring, but once our first human data comes in, we will be looking to scale up our team quite a bit! If that's something you're interested in, drop me a line, email is in my bio.

At our school, pong was given as a freebie template for building your final project in the digital design course! A great time, ended up enjoy it so much I TAed it till graduation.

Do we know the terms of this raise? I wouldn't be surprised if it's released in tranches based on KPIs.

That's fascinating! Insulin sensitivity is an incredibly complex mechanism, and is always in flux. Hyperinsulemic clamps are one way to measure the boundaries of insulin sensitivity, but doesn't really help you understand the day to day individual dynamics very well. Like I said, drugs are _not_ our area of expertise, so I can't comment much more than that.

State of the art treatment these days (at least in artificial pancreas land, where I spend all of my time) is more focused on sensitivity-agnostic treatment; figuring out the insulin sensitivity dynamically rather than trying to control the sensitivity itself (it's much more multifaceted than a few hormones, we're talking hundreds of possible ways it can change).

Also, and I'm sure the authors would acknowledge this, n=32 is enough to demonstrate a possible effect, but not nearly enough to show this effect in the population at large. You usually need n~5000 or more for such effects to be shown generally, though the actual number depends a lot on the drug and what you're trying to treat.

We're conducting our first human study at a world renowned US diabetes center soon, I'll see what our PI thinks of this work. Thanks for sharing.

We're not really in the business of drugs/supplements, so unfortunately I can't comment on the efficacy of ALCAR and other drugs on the treatment of diabetes. If you have any papers on the topic, feel free to email me (linked on my profile) and I would be happy to discuss when there's time!

Also, as much as I wish we had any sort of platform through idealmedtech.com, our traffic is quite low, mostly from investors and people pitching SaaS products

I think it's clear from the pandemic that stock values do not correlate with the economy.

I remember a friend in my dorm playing Lugaru back in 2008, it was such a fantastic game back for the time! I still vividly remember the looping soundtrack and the fight effects. Thanks to the team for making sure this little piece of digital history is remembered!

My use case is closer to RDP, but as I understand it, you cannot host RDP through a firewall without Windows Server, which I don't have (plus, I don't have much interest in opening a port into my home network). I'm looking for something that is always enabled, so I can access a machine while it's unattended (so Zoom and the like are out, since there's a 1 hour limit and you need to click a link).

Maybe Chrome Remote Desktop? Or RDP, but using a VPN of some sort? I know there's a little-hassle tunnelling setup you can use a la WireGuard that's been on HN, but I can't remember the name