achileas

Except there weren't really any mass harassment, rape and death threat, and firing campaigns being coordinated against ordinary people for not sufficiently mourning someone. Most of the "cancel culture" stuff was overblown nonsense, the few real events were against massive public figures credibly accused of heinous things like Weinstein.

Pretending this is in any way equivalent betrays either an intense naivete or a supporter of this pre-pogrom behavior.

They didn't claim that there were any, just that AGI isn’t a necessary requirement for an application to be world-changing.

I've used Promptfoo for client projects and really enjoyed it once I got my head around it.

This makes me want to build agents in Elixir now

Agents don't typically, and any LLM they're calling is likely hosted remotely.

That doesn't negate what OP was saying at all, the better support in Python isn't for "running ML/AI code" but in things like agent frameworks, observability tools, SDKs, etc. None of which directly run AI code but are still helpful/necessary and for the most part better represented (and supported) in the Python world, although that seems like it's slowly changing.

Maybe I've aged out of it or something but the userbase has gotten so much more annoying (and was already annoying when I finally started using it around 2010) over the last several years, that I've almost entirely quit it. I used to mod large subs, I started a company with a group of guys I met on Reddit (and we still have our own group chat), and its just completely lost all appeal to me. It's like every user is the annoying college freshman that thinks they're an expert on all things because of half a semester of intro physics.

And destroying the rest of their economy to do it

It’s like a whole post of all the things you’re not supposed to do with Python, nice.

Also years for both for me (I probably read and watched them in 2003), but your impression of the two is the exact same as mine.

Battle Royale was a novel before it was a movie, and it's a much better novel than movie.

No, doping is a specific set of techniques that use proteins (EPO) or specially extracted blood to boost the red blood cell count in the blood, increasing oxygen carrying capacity.

Anything else is just marketing.

tbf I've never seen expiration dates on fresh produce in the US. Maybe some packaged salads and things (and it's never accurate, of course), but nothing fresh.

And of course, not just frequency of incoming action potentials, but processes within the receiving cell, in the cell membrane, at the site of the synapse, and between the cell and any supporting cells (astrocytes and glia).

It's also not just frequency, but "shape" (for lack of a better word) of incoming inputs that matters, as such there is a very wide variety of spiking patterns that certain cells exhibit, like chopper cells.

We don't really know much about how memories are formed, the short-term and long-term divisions come from cognitive psychology and we don't have (from what I remember from my neuro days, which may be out of date by now but I am friends with several neuroscientists and I'm not sure it's out of date yet) a strong idea of what the biological correlates of each are, or even if they arise via different processes or what.

None of which are causal associations. Given the millennia-long history of cannabis use to self-medicate, and lack of evidence (not without trying!) for a biological mechanism of any of this, it’s probably safe to assume this is largely people with an issue (or a proto-issue) self-medicating.

I have a graduate degree in neuroscience, worked with colleagues who focused on psychopharmacology for their research, and many of my friends and neighbors are biologists of various stripes, including still-active neuroscientists, as well as epidemiologists, and clinicians. They all agree cannabis is fairly harmless, and would outright laugh you out of the room if you compared its negative effects (either in the individual or to society) to alcohol.

Clinicians aren’t the ones to go to for harms anyways, they’re largely not doing the research at any level.

I can't even fathom the utter despair of losing a child, and then beyond that to be hounded, "investigated," and arrested for it, constantly forced to relive the worst moment of your life on top of the unbelievable grief of losing your child.

If I never have to hear from finance dorks pretending to be engineers tarnishing the title "bro" right this second it wouldn't be soon enough.

This gives me heart. I like writing about technical things, but I also like writing about personal things, concerts I went to, whatever. I'm a whole person, and I never liked the pressure (mostly from social media) to build your "brand" around one genre or style of writing. For me, my site is a personal one where I post about things I'm interested in. Ham radio, machine learning, my travels, pay phones, whatever. Maybe less useful for a reader or audience building but...I just like to write and share things.

I think it just depends on how you're able to access and use your notes. For me, backlinks were a life-saver because it's important to model relationships between individual notes. It's really helpful for synthesizing knowledge and creating output, or just exploration to find related ideas, etc. Backlinks also help with breadcrumbs, so if my memory fails (frequently) or my search fails (more often than I'd like), I can usually find what I was looking for by navigating backlinks or using hierarchical breadcrumbs (all generated automatically with backlinks in a defined property). So I get a bunch of bottom-up organization for free with minimal effort, and for me it's pretty helpful. Plain textfiles I have trouble resurfacing knowledge very quickly, but YMMV, it's really personal.

This is something that Obsidian, Tana, Roam, LogSeq, and a few others actually excel at. I try to keep my organization minimal or at least static, and backlinks (plus templates, or in Tana supertags) + good search (of all that I've used, Obsidian has the best) help to take organization off your plate and make it more fluid. I've also built a pretty rudimentary RAG MVP to make finding information even easier, but don't use it too much right now.

Obsidian isn’t opinionated on the system you use, but it’s trivial to implement a zettelkasten system - I’ve been using a modified one (inspired by Andy Matuschak’s evergreen notes [1]) for several years now.

[1] https://notes.andymatuschak.org/%C2%A7Note-writing_systems

She and her colleagues calculated that the large language model BLOOM emitted greenhouse gases equivalent to 19 kilograms of CO2 per day of use, or the amount generated by driving 49 miles in an average gas-powered car.

If you go on to read the paper, you discover this is an outright lie. These figures are for training, not usage.

ZH is like a less-respectable OAN

It’s also very much not just about the neuronal synapses.

That’s not true at all. The most interesting research into components of consciousness involve astrocytes, glia, and formerly assumed “helper cells” throughout the brain. It isn’t “all synapses.”

Source: neuroscience MS (dropped out after defending to become a software engineer)

Not OP, but old dentist made and sold these. It is a custom-fit mouthpiece that forces you to keep your lower jaw forward while you sleep, preventing airway collapse. It’s the same thing you’re taught to do for CPR to force the airway open.

My old dentist made and sold these, but the oral appliance is a custom-fit mouthpiece that forces you to keep your lower jaw forward while you sleep, preventing airway collapse.

Form my own POV you’re correct. The market is good, salary offers I get are normal, everything I see looks like the pre-2020 market to me. A lot of these reddit CS subs are full of kids who are either still in college or graduated into the hiring frenzy, so their perceptions are pretty skewed and they tend to catastrophize.

Plus, one weird lowball offer and a weird “explanation” from some recruiter isn’t indicative of a trend.