sometimes they randomly choose the ugliest possible way to do pattern matching, eg multiple blocks of nested "if let" instead of a "match", or a "match" instead of a single "if let"

otherwise, works great; much easier to un-vibe the code compared to eg python

(gpt 5.* in codex/sonnet 4.5 in cc/glm 4.6)

I believe ADHD might be neutral/beneficial in some environments, and I also am not a big fan of labels; if it has a non-negative impact on your life -- sure, you don't need the meds

but

Until I started doing stims, I was regularly forgetting about food/water until I could barely move, and at one occasion I procrastinated from replying to an email with the offer of my life until the deadline was over. There's a lot that you can explain with "not motivated enough", but these just don't make any sense, right?

You do need some sort of a label to prescribe controlled substances. Even if you forget about the meds, just knowing that it's a pretty common thing and you're not alone is pretty helpful.

I mean, it does take less space (compared to having one per window)

Only if you have two or more windows vertically stacked, which isn’t what you’re likely to be doing when you want to maximize vertical space.

but if they're stacked horizontally, buttons don't fit

I really like how Kagi does it - summary only appears if you finish your query with a question mark, or if you press "q" after loading the results

> Are they "decisively" choosing it

I'm pretty sure chatgpt and perplexity are "opt-in" in this sense

> nothing to do with language models

how do you think translation works?

reminded me of "tensor is a bunch of numbers that transform in a certain way"; this should be illegal to teach, especially in physics

> not some true disorder

there's a tool on your desk that might help you solve your problem; what does it matter if the problem is an "appropriate response of your body"? so is pain/anxiety/diarrhea

> stems from living a life that is not true to myself and due to countless failed attempts to be someone I cannot never be.

If this is their mindset, they might benefit from CBT more than medication.

I'm not against SSRI at all. But after taking them for a few months in my 20's, and experiencing how terrible the withdrawal symptoms are when stopping, I'd be very hesitant to ever start up on them again. I remember having to open up the lowest dose pill capsule and splitting the dose into very tiny increments to be able to wean off completely.

I'm the GP of this comment chain. I actually did a 16 week long trial of CBT with a professional where my symptoms were tracked during each week. I actually ended the 16 weeks worse than before I started.

I later learned that CBT can have that effect on people with ADHD, so I attribute that to being a possible explanation.

I still do therapy, but honestly, I think it's a waste of time and money. I predominantly do it for cathartic purposes and so other people/professionals will stop recommending it.

> terrible the withdrawal symptoms are when stopping

I tried one when I was in my early 20s too. I swore I would never take one again. Withdrawals weren't my issue really. It was the clarity I gained after getting off. I realized how awful I was to people around me. I had such blunted emotions, that I basically became devoid of empathy. I also learned that I needed high levels of anxiety to function, which the medication took away from me.

Are you taking something for ADHD?

I think I had a somewhat similar experience: before antidepressants, I was practically non-verbal; later, I had the typical brain fog/emotional blunting -- still unable to communicate with the therapist; after adjusting to meds, by the time that I can put the problem into a coherent sentence, I can usually fix it anyway, so what is the therapist for? Is it really just rubber duck debugging, where the duck has a medical degree?

One thing I wasn't able to figure out was ADHD -- I didn't really believe I had it, as my psychiatrist was sure it's not the problem -- everything was explained with depression/anxiety/geniune lack of motivation. Now that I finally have a prescription for stims, seems like it was the other way around: most of the time I would have a feedback loop where my typical fuckups would trigger a downward spiral, and it's almost not a thing anymore; being able to get shit done is just generally such a pleasant experience!

TLDR My depression also wasn't a "true depression" in a sense, but you still need to fix some underlying issue, which is much easier when your thoughts are not all "I'm a failure" shaped

for me, medication was closer to therapy than what I imagined (magic chemical that makes depression disappear) -- a drug changing my thought process (long term and on a deep level, unlike recreational drugs) helped me see myself as something malleable, instead of the idea of somewhat rigid "true self" I had before; from there it was much easier to start doing the DIY equivalent of CBT, as I stopped considering some of the feelings/ideas to be somehow inherent to my personality

> so much slower

unless you're working on 10 features in parallel with smaller models in best-of-4 groups

counterpoint: I want AI in firefox

there's so much stuff that could get much better if they invested more in AI features -- tab grouping, translation, ad blockers; why are people so triggered? because it might end up being bad?

Based on the main demographic that normally jumps to Firefox, it would at least be a good idea for them to make the features opt in rather than opt out. Most of this stuff is on by default.

I'm curious about what you think AI would do for those features? I've never had issues with ads after just installing stock uBlock Origin, and local translation is already available and works great for me across the web. I'm not sure what AI would do for tab grouping. Are you envisioning having 100+ tabs and then telling an AI assistant to sort it all out for you?

> good idea for them to make the features opt in

true, but that's not the same as "nobody wants AI"; them fucking it up once shouldn't mean that they should drop the entire idea; by this logic they should stop making browsers altogether

the only thing actually pissing me off about this is the selection popup, everything else is just yet another (useless) button in a menu

> local translation is already available

and it's already "AI"

> I'm not sure what AI would do for tab grouping ...

they already have a "suggest more tabs" thing that does exactly this using a local model; it's insanely slow, low quality, doesn't use gpu, but the idea is nice. I'm hoping they will continue working on it, as managing tabs across multiple windows is not exactly what I want to spend my time on, and it's exactly the type of a problem LLMs are perfect for

> never had issues with ads

it could always be better; there's always a certain amount of sites that either don't work at all or still show ads, and I imagine you could use a local VLM to hide elements based on the actual rendered result; same with cookie banners, annoying sign-in prompts, and so on -- how is this not the coolest thing ever? as long as you have like 8gb (v/u)ram, you could immediately unshittify most of the web

AI is unfortunately an umbrella term which people can project what they want on to. For reasonable people like yourself it's a tool to accomplish tasks. While for someone else it might be CoPilot showing up frequently and they don't disable or turn off the notifications and just get continually frustrated.

so why not give mozilla the benefit of the doubt? why assume that it's necessarily gonna be a slopfest? I'm not exactly optimistic, but it could be good, I do want those features, and I'm not gonna switch to something like chrome

by this logic sockets are also non-unix

I mean, that's not exactly wrong...