1,060 karma · joined November 15, 2022
> n. the dread of finally pursuing a lifelong dream, which requires you to put your true abilities out there to be tested on the open savannah, no longer protected inside the terrarium of hopes and delusions that you started up in kindergarten and kept sealed as long as you could.
God I hope I don't fuck it up. But at least I tried.
I just use kysely for my sqlite-in-browser project.
https://github.com/open-spaced-repetition/fsrs4anki/wiki/ABC...
If you ever feel like moving beyond the Apple market you might look into https://github.com/vlcn-io/cr-sqlite I've been using it in my own Anki-clone and am very happy with it.
> A free, open source, local-first, spaced repetition system that works offline, has p2p syncing, plugins, and first class support for collaboration. It's GitHub/Reddit for flashcards.
https://github.com/AlexErrant/Pentive/
All the "hard" tech has been more or less built out... all that's left is drawing the rest of the owl >_>
I would be curious if there's been an effort to get tree-sitter working on the web.
Slightly off topic; is anyone aware of any Lezer grammar for regex? I've not been able to find any in the FOSS world. I suspect Regex101 has one, bit it's sadly closed source.
You fine-tune the LLM to spit out RPG skills based on the player's chosen class. (Dwarven princess = summon dwarf minions, has a rare bloodline magic, whatever). The classes could get _real_ weird, like a school bus character could summon school children minions and skill up to flying a la The Magic School Bus. (Or that could be a prestige class haha.)
A second model would match the RPG skills to a known/finite set of game mechanics supported by your game. This is ultimately a simple classifier - not really an LLM (though the skills on the input side are language, okay pedants). Determining the set of game mechanics to support shouldn't be too hard - how many different ways of doing damage and (de)buffs could there be? (Okay there can be a shit-ton.) Balancing the game could be done live via reinforcement learning. To constrain the search space, it'd follow the old Zelda format of limiting you to one room/screen until the enemy is defeated (or like Binding of Isaac for those too young.)
It'd have to use 2d/sprite based/placeholder graphics because I don't think the rigging/animation AI is currently good enough... but bad graphics hasn't stopped fun gameplay before. Imagine Charizard vs Death Star. Good luck with the DMCAs.
Another problem is that with simple graphics and literally infinitely variable enemies, there's no learning on the player's side so there's no real "skill" accumulation to improve on for the next run. There would be a lot of reading on the player's part if you decide to just show a stats-list for an enemy type. I suppose you could turn it into a zombie game where new enemies are slowly introduced in a horde of other enemies the player's familiar with.
> A paper from 2020 [0] showed that Cranelift was an order of magnitude faster than LLVM, while producing code that was approximately twice as slow on some benchmarks.
---
On a related note, I love posting this study https://jamanetwork.com/journals/jamanetworkopen/fullarticle...
Look at Figure 2C in particular. The hazard ratio of smoking is 1.41. The hazard ratio between "Low" (<25th percentile) and "Below Average" (25th-49th percentile) cardiovascular health is 1.95.
In other words, it is better to be a smoker than in the lower 25% of non-exercisers. By far. It's not even close.
https://github.com/asciinema/asciinema-player/blob/develop/p...
https://news.ycombinator.com/item?id=29387761
The Rust frontend frameworks have been creeping up the perf charts, though.
Tab/Window hunting via CtrlTab/AltShiftA is... not great. A goal of mine is to never AltTab.
> We used Cox proportional hazards models to estimate hazard ratios... between self-reported hearing aid use status (hearing loss with or without hearing aids) at baseline and risk of dementia
The top-level article discusses a randomized controlled study and therefore can make a much stronger claim of causality:
> The researchers randomly assigned participants to one of two interventions. About half received hearing aids and instruction in how to use them. The other half were assigned to a health education program focused on promoting healthy aging. Both groups received follow-up visits every six months to reinforce the training.
Shameless plug - I'm building https://github.com/AlexErrant/Pentive which is basically GitHub/Reddit for flashcards. Very much pre-product and a WIP, though the offline client proof of concept is done.
> Extreme cardiorespiratory fitness (≥2 SDs above the mean for age and sex) was associated with the lowest risk-adjusted all-cause mortality compared with all other performance groups.
Sure your cardio might be "good" now, but see Table 2 for what happens to you over time (with and without training).
Put more plainly https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6153509/
> Individuals with higher cardiorespiratory fitness (CRF) showed lower risks of all-cause, CVD and cancer mortality; those with higher grip strength (GS) had lower all-cause mortality. All-cause and CVD mortality risk was lowest in adults with both higher CRF and higher. Improving both CRF and muscle strength, as opposed to either of the two alone, may be the most effective behavioral strategy to reduce all-cause and cardiovascular mortality risk.
> Mitochondrial dynamics, including continuous biogenesis, fusion, fission, and autophagy, are crucial to maintain mitochondrial integrity, distribution, size, and function, and play an important role in cardiovascular homeostasis. Cardiovascular health improves with aerobic exercise, a well-recognized non-pharmaceutical intervention for both healthy and ill individuals that reduces overall cardiovascular disease (CVD) mortality. Increasing evidence shows that aerobic exercise can effectively regulate the coordinated circulation of mitochondrial dynamics, thus inhibiting CVD development. This review aims to illustrate the benefits of aerobic exercise in prevention and treatment of cardiovascular disease by modulating mitochondrial function.
Attia is a fan of "zone 2" training in particular because it trains mitochondria to burn fat, which leads to "metabolic flexibility". https://pubmed.ncbi.nlm.nih.gov/28467922
> Metabolic flexibility is the ability to respond or adapt to conditional changes in metabolic demand. This broad concept has been propagated to explain insulin resistance and mechanisms governing fuel selection between glucose and fatty acids, highlighting the metabolic inflexibility of obesity and type 2 diabetes. In parallel, contemporary exercise physiology research has helped to identify potential mechanisms underlying altered fuel metabolism in obesity and diabetes.
Weightlifting, if anything, depletes muscle glycogen, a sugar, so you aren't really training your fat-burning.
That article was written in July, and in the intervening time, Chrome's announced an intent to ship "Multiple Readers and Writers in File System Access API" aka `readwrite-unsafe` [0] in v121 (stable in 2 weeks) which could help improve SQLite's performance even more.
[0] https://chromestatus.com/feature/5172892632875008
Edit: It looks like wa-sqlite already has a prototype VFS taking advantage of the above feature https://github.com/rhashimoto/wa-sqlite/discussions/116
https://www.majorgeeks.com/content/page/youtube_double_tap.h...
> our finding showed a significant increase in breast cancer mortality rate in the world during the past 25 years ... Statistical analysis showed a significant increase for breast cancer mortality rate in all super regions, except for High-income super region. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6745227/
and for women under 40:
> Female breast cancer mortality rates have stopped declining in women younger than 40 years, ending a trend that existed from 1987 to 2010. Conversely, mortality rates have continued to decline in women aged 40–79 years. https://pubs.rsna.org/doi/full/10.1148/radiol.2021203476
I disagree that rote memorization limits innovation in the medical field. Biology has complex problems, and "For every complex problem there is an answer that is clear, simple, and wrong". There's a reason why so many clinical trials fail, and it's not for a lack of technical innovation.
I feel like the medical _system_ as a whole needs systematic/organizational disruption, not technological disruption. For example, CGMs are a technology that have been around for decades and only now are we seeing usage in non-diabetic populations. The system (they require a prescription) is preventing the innovation.
Unfortunately, hard disagree. The best way to learn a Lisp is to write Lisp. You can look at code all you want, but you won't get used to the syntax/parens until you start writing it. Same applies to non-Lisps.
> I think a lot of people program in Python but programming in Python is not their main occupation. These people struggle to accumulate knowledge, because they may encounter concepts or APIs only from time to time, and by that time they already forgot.
This is a great point. Static typing is great for API discovery, which Python sadly lacks. However, I wonder how much LLMs are eating this "low-code" market. Also, if Python is non-critical to their jobs, I suspect that they won't feel motivated to explicitly study Python over the long term. Doing spaced repetition is famously hard, and non-motivated people will likely give up and just do "JIT learning".
> Also, I want to cover some "unorthodox" topics for which spaced repetition may be specially well suited. For example, I'm building a deck called "A tour of the standard library", where you can learn all the modules that are available, just so you know what is out there, without going deep into it.
FWIW I'm learning Rust, and to "learn what's out there" I've just been binging random Rust videos on Youtube at 2x over lunch. Granted I'm an experienced dev and I have an intuition as to what "should" be out there, so perhaps this story is of limited use.
Haha, monads still haven't clicked for me despite knowing F# and Clojure pretty well. I think of them as "functors that you can flatten" (and functors are "things you can `map` over", and `map` is "a structure-preserving transformation"). I know monads syntactically, but not intuitively. Despite this I still manage to work with monads pretty well - sometimes rote knowledge is enough. Maybe if my work involved discovering new monads I'd feel like I'd know them better. (I just see the same option/list/result(ish)/generator-style monads over and over.)