samuraixp

Stupid question but? Fake red LEDs that have battery backup in every security camera?

Seems cheap and effective.

Pre AI today it seems like we’re in some weird place where IQ scores are dropping in ways that don’t match the Flynn effect and people like Bryan Caplan argue that education is just virtue signalling to gain access to higher status and paying bullshit jobs (David Graeber). Everyone here talks about how real that is in the world of whiteboard interviews and rest and vesters. It seems pretty true elsewhere in the world of white collar nepotism.

What are we really educating kids for these days? To have advantage over other kids because we have no fair meritocratic way to allocate resources or meaning in society?

Won’t AI just make this infinitely worse.

Like, one vision is more teachers and more students learning better, and another is less teachers and more baby sitters and lower government budgets for education leaving students with the equivalent of an automated telephone answering service menu instead of a real human call centre?

I agree, except all those kids just go on to work in an Amazon warehouse driving a forklift anyway.

We need to start valuing human intellect more otherwise it will simply become a thing we have outsourced to machines like washing clothes and transporting our bodies.

Pretty cool, I had a little chat in german and then the last translate button wouldn't work, the chat disappeared after I clicked around the rest of the interface. Main thing that seems missing is there is no feedback on the grammar and spelling or things that should be said, so perhaps you need a second model + prompt to analyse the inputs of the user in relation to the conversation and critique it as a language teacher in terms of spelling, grammar and contextual relevance, suggesting an alternative that the teacher would have said?

Very recent father who asked to have his son seeded after a unplanned c section here. Iv followed a lot of recent micro biome research and it does seem weird to think there is any downside to doing this given that a vaginal birth contains a lot more of the same seeding microbes. GBS seems overblown but you could still pre test for that earlier in pregnancy if you want to play it that way.

Baby Koalas literally cannot digest eucalyptus leaves unless they get the microbes from eating their mothers faeces after birth. Mammalian microbe seeding during birth, is likely critical to our evolutionary history.

This looks awesome. The state of virtualization on Apple Silicon right now is a bit painful. If this really does provide semi performant x86 Linux emulation and there was a Vagrant Provider plugin then this would easily be the defacto tool for dev / testing VMs on macOS.

https://github.com/hashicorp/vagrant/tree/main/plugins/provi...

Same probably applies to hashicorp Packer.

Every tool I have tried like Parallels and VMWare Fusion 13 Pro says that nested virtualization for windows is not possible. If this is possible with Orb at some point i’d pay for it.

I often want to test a Windows VM with Docker installed into WSL2 and this becomes a nightmare now on Apple Silicon.

Also I wonder what possibility this opens up for an improved toolchain to develop stuff like Asahi Linux by bridging macOS native tooling and Linux emulation to write and test code without rebooting or using two machines.

Amazing work. Why can’t Apple, Vzmware, parallels or someone else do all of this when a single developer can. Sad.

So happy for the author that they discovered the solution to their chronic health issues. Some estimates suggest that 10% of the population are suffering undiagnosed health issues. Thats a lot of people whos quality of life is being potentially severely disrupted because the health system is not designed to provide deep dive diagnostics. Surely the productivity gains in solving these health issues would easily pay for themselves. I hope this is an area where AI will truely shine by providing an effective stepwise differential diagnosis for anyones set of vague symptoms and conflicting biomarkers.

Wow, super happy for you. Just to add to this I will say that I was mis-diagnosed with an adrenal lesion which turned out to be incidentaloma. I had an adrenal gland removed however the treatment was non curative.

At the time I asked why they weren't suggesting adrenal venous sampling and was looked at like I was an idiot by the resident doctor. Turns out she was the idiot. A specialist later told me they were probably too unskilled to do the adrenal sampling.

Also the cyp11 b1/b2 PCR test has existed for ages and is in the literature. Years later I had some DNA testing and discovered I have an even rarer version of PA called Familial Type 1 caused by a non stop mutation on CYP11b1/b2, hence why the adrenalectomy did nothing.

While its rarer, the test is only a few hundred dollars. To others... a pro tip: get a pre-genetic screening test even if you aren't having kids, the cost is cheap and the result might be life changing to you TODAY.

I feel an order of magnitude better since silencing the over production of that hormone and my doctor tells me that anecdotally practitioners report patients coming out of curative surgery for nodules or on dexamethasone for type 1 often report feeling like a cloud is lifted from their mind.... I can confirm this is true.

Sadly you really do need to, push, research and be your own advocate to get the best outcomes.

I find this really annoying. There's also no concept of a package like Helm Charts which means you can distribute all your images on Docker Hub but then when it comes to compose files and configurations with templating your up @$%^ creek.

I was hoping something like https://dagger.io/ might solve this but they currently don't support compose files, or maybe https://www.acorn.io/ ?

It does feel like there is a legitimate space for a multi-container service application standard that isn't just kubernetes, unless perhaps things like k3d and docker desktop end up so good at running tiny k8s that docker compose dies.

I have been getting pretty high VOC readings in my bedroom on a uHoo, Airthings and now even the Dyson HP09 shows some lower levels. So now I am confident there is something being detected. Any advice on how to debug what is causing it and what VOCs it might actually be?

The weird thing is it seems to mostly spike at night time and even more so on the weekends.

I have a few hypothesis, such as: - furniture offgassing - us being in the bedroom - gas cooking in the apartment building somehow settling into our ground floor apartment at night? - it seems to be worse if we have the aircon on - it really only seems to properly dissapate when the windows are open and theres good breeze and there seems to be a correlation between that and the sun coming up?

Is any of this content duplicate with the book? Should I just go read that?

Love the "tour of duty" idea. Also, Chicken Salt for the win!

Damn, William Gibson called this one with Idoru.

Funny how the idea of a virtual pop star in the 90s barely made sense but now post influencer it’s like… yup totally.

Surely this is the beginning of providing a side loading option. Perhaps if they enforce the purchase of a paid developer account they can get their pound of flesh from the users on their way out of the store. It would be ironic if iPhone users paid Apple 100 USD a year to side load to play Fortnite etc...

Dataset and interface for training faces on your very own, low cost, highly parallel, CUDAless biological neural network.

This happened to me ages ago and I was also unable to recover my accounts but recently I tried again, and I was able to recover 1 of my 2 lost accounts. The second one is pending. It seems Google has changed the process and after a semi-failed attempt to recover they actually pass it to a human. All my previous attempts ended in a stone walled error "We are unable to verify it is you", even when I was able to provide things like old numbers, old passwords or verification codes from backup emails.

Maybe try again. :)

One thing to note is that I was unable to access an old account for so long that it no longer has any old email in it.

I use http://papersapp.com/ for handling PDF papers from arxiv.org etc.

It seems they have merged with readcube who want a yearly subscription however I was still able to buy a life time license to papers app a few months ago. I guess support will slowly disappear but for now its working great, with PDF sync via dropbox onto my iPad or iPhone, the ability to draw and take notes and search as well as download new papers. For your own PDFs I guess this would still work pretty well since you can search and annotate, create smart groups etc.

Awesome, so what qualifications do your Data Science team members have and would you hire someone without a degree? Or what if they had a data science / ML / AI MOOC like Udacity or Coursera? And do you have any advice for OP or others who don't have degrees?

me 2 hahahah Go Sal!

Does anyone have a working version of phantomjs 2.1.1 with ghostdriver that connects to selenium grid?

I tried this one for example but it just crashes randomly when the jobs are running: https://github.com/AkeemMcLennon/docker-selenium-node-phanto...

The latest 2.5.0 beta doesnt have the updated ghostdriver yet.

How much ram does chrome --headless use (true headless not XvfB), because I want to run about 200 of these in parallel and keeping the ram down is important obviously.

Does anyone have any suggestions? I figure by using selenium grid I can run them in docker containers and even on multiple hardware machines.

Currently, I am running about 50 in parallel with Magellan https://github.com/TestArmada/magellan

But its hitting some kind of system limits or sync issues because each process is started before run time rather than ready and waiting in a pool.

Any help on running 200 headless browsers via selenium grid would be greatly appreciated. I think the current selenium chrome Xvbf docker container uses like 300mb+ per instance so that would be 60gb of ram which is a fair bit.