Хороший допис від Карпати на Хвіттері про сучасний стан розробки із LLM-ками
(хороший у тому сенсі, що ємний, правдивий та без надлишкового ШІ-гайпу)

Посилання на оригінал залишу внизу посту, а основні тези ось:

Coding workflow:

[...] like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in December.

IDEs/agent swarms/fallability:

The models definitely still make mistakes and if you have any code you actually care about I would watch them like a hawk [...]

The most common category is that the models make wrong assumptions on your behalf and just run along with them without checking. [...] They also really like to overcomplicate code and APIs, they bloat abstractions, they don't clean up dead code after themselves, etc. They will implement an inefficient, bloated, brittle construction over 1000 lines of code and it's up to you to be like "umm couldn't you just do this instead?" and they will be like "of course!" and immediately cut it down to 100 lines.

Tenacity:

They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later.

Speedups:

Certainly I feel net way faster at what I was going to do, but the main effect is that I do a lot more than I was going to do because 1) I can code up all kinds of things that just wouldn't have been worth coding before and 2) I can approach code that I couldn't work on before because of knowledge/skill issue. So certainly it's speedup, but it's possibly a lot more an expansion.

Leverage:

LLMs are exceptionally good at looping until they meet specific goals [...]

Write the naive algorithm that is very likely correct first, then ask it to optimize it while preserving correctness. Change your approach from imperative to declarative to get the agents looping longer and gain leverage.

Fun:

I didn't anticipate that with agents programming feels *more* fun because a lot of the fill in the blanks drudgery is removed and what remains is the creative part.

[...] I experience a lot more courage because there's almost always a way to work hand in hand with it to make some positive progress. I have seen the opposite sentiment from other people too; LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building.

Slopacolypse:

I am bracing for 2026 as the year of the slopacolypse across all of github, substack, arxiv, X/instagram, and generally all digital media. We're also going to see a lot more AI hype productivity theater (is that even possible?), on the side of actual, real improvements.

https://x.com/karpathy/status/2015883857489522876

людина використовує свободу волі

by sranymiyab

(Додаю витиск із статті (себто сам її кінець) — НЕ МЕДИЧНА ПОРАДА)


MY SPECIFIC RECOMMENDATIONS FOR YOU TO DO A.S.A.P:

- Go to a pharmacy, buy the following supplements over-the-counter, in whatever form you like: (I like the easy-to-swallow gel capsules)

Vitamin D

- 🌱 By default, Vitamin D supplements are derived from… (quick web search)… the grease in sheep's wool? Huh. Also fish liver oil. Anyway, if you're vegan, make sure your bottle specifically says "vegan" or "from lichen/mushrooms". (If you're vegetarian, the sheep's-wool Vitamin D is fine, they don't kill the sheep for it.)


Omega-3 where EPA is ~60% of the Omega-3 total.
For example, my 500mg Omega-3 capsules have 300mg EPA, 200mg DHA.

- 🌱 By default, Omega-3 supplements come from fish. If you're veg(etari)?an, there are plant-based sources of Omega-3, but look carefully: most vegan Omega-3 supplements provide more DHA than EPA, which the above studies suggest fully cancel out the antidepressant effect. Double check the nutritional label to make sure it's ≥60% EPA.


Then, every day:

Take ~5000 IU of Vitamin D
- ⚠️ be cautious if you have kidney stones, or are on medications that could interact with Vitamin D. "ask your doctor".
- 4,000 IU is the "official maximum safe dose", if you understandably don't trust a random internet blogger, even though she cited peer-reviewed sources.
- 10,000 IU if you're feeling daring / have darker skin / live in less sunny climates.
- bonus: may improve immune response to Covid & influenza?

Take ~1500 mg of ≥60%-EPA Omega-3
- ⚠️ be cautious if you're on blood thinners, or other medications that could interact with Omega-3. again, "ask your doctor".
- bonus: may improve cognition?

(Don't quit your existing antidepressants if they're net-positive for you!)
- you may also want to ask your doctor about Amitriptyline, or those other best-effect-size antidepressants.


Can you get these doses of Vitamin D & Omega-3 through whole foods alone, no supplements?

Probably, but it'd be expensive & tedious: you'd have to eat something like 2,000 calories of farmed salmon a day to get 5,000 IU/day of Vitamin D. As for Omega-3, eating mostly oily fishes would get you >1000mg of Omega-3, but they'd be more DHA than EPA, which the above studies suggest would cancel out the antidepressant effects.

Re: Vitamin D and Omega-3 have a larger effect on depression than antidepressants

Please do not take 5000mg/day of Vitamin D. The author confuses IU and mg which is very dangerous.

r1ch, 2 hours ago

В оригінальній статті (уже) та у витиску вище інфа коректна

але майте на увазі

поїхали в лондон щоб зібрати гроші

на нрк для любарту. кожна з нас відкрила банку щоб зібрати якнайбільше. хто збере найбільшу суму за 3 дні, розіграє локально-лондонський приз серед своїх донатерів

будь ласочка репотік і донатік
банка: https://send.monobank.ua/jar/28u6sALsxe
пейпал: [email protected]

Відкриті курси кафедри математики КАУ (онлайн формат), весняний семестр 2026 року
(для бакалаврів і магістрів університетів України і не лише)

1) Групи Лі (Максименко), по четвергах в 14:00.
Перша лекція - 5 лютого 2026 року

2) Функціональний аналіз (Островський), по середах в 11:00.
Перша лекція - 4 лютого 2026 року

3) Симетрії диференціальних рівнянь (Бойко), по четвергах в 11:00.
Перша лекція - 5 лютого 2026 року

4) Вступ до машинного навчання (Тимчишин), по п'ятницях в 14:30.
Перша лекція - 6 лютого 2026 року

Якщо Вас зацікавили наведені вище курси, пишіть https://t.iss.one/m_mokka (Олена Мочук) щодо приєднання до відповідних груп.
Студенти, що успішно здадуть іспити, отримають відповідні сертифікати про прослухані курси.

Telegram

привіт

я наразі шукаю тимчасове місце проживання для своєї подруги. подруга була ізбита своїми батьками до крові і... блять я не знаю що ще дійсно казати

сама подруга проживає між Тальним та Уманем

це не якийсь байт чи прікол, це серьйозно

благаю хоч чимось допомочи

дякую зараніше за допомогу...


p.s. ось ще банка якщо ви хочете ще підтримати фінансово

https://send.monobank.ua/jar/82mUjH9aoK

контакт зі мною - @non0x3

#article
The complexity of misinformation extends beyond virus and warfare analogies

main reason for posting this is that image goes hard