Аналитика Telegram-канала
Подписчики
1 663
Охват за 24 ч
181
ER за 24 ч
10,88%
Публикаций
35
Динамика подписчиков
7 дней
30 дней
6 месяцев
Больше
14 ноя17 ноя20 ноя
Показаны последние 7 дней
Больше истории →
Охваты публикаций
15.1115.1117.1118.11
46
Средний охват
57
Лучший результат
Показаны 4 публикации — последние в канале
Больше статистики →
Упоминания
39
Упоминаний
0
Из них репостов
22
Каналов
0
Упоминаний за 90 дней
0
Из них репостов
0
Каналов за 90 дней
Последние посты
3 из 35 2 из 35Everyone I know is cashing out stablecoins this week to buy ONE stock
Anthropic goes public on Thursday with $ANT
It’s set to be the second biggest IPO in human history
So why is this IPO so attractive to everyone?
First and most obvious, Antrophic is one of the leaders in the AI industry
But a title alone offers no guarantee, just think of Minimax Group, which is currently trading at $14, but after launching at $30, it hit an all-time high of $150 (5x)
BUT we’re dealing with a company that’s larger and more well-known worldwide, whose product is used by almost everyone you know
Let's move on to the statistics:
> Revenue for the past quarter was $17B (12 times more than a year ago)
> 11,900+ customers paying $100,000+ per year (9 times more than a year ago)
> ~250 customers pay $10M+ per year
> ~1,900 pay $1M+
> Revenue per employee is $6.6 million
The deal itself is simply massive..
Full read: https://x.com/0x_kaize/status/2123021605546979402
ещё
56
0
1 18 ноя, 14:57
All code is slop. Except mine.
That’s the title of Lauren Tan’s talk at the AI Engineer Code Summit, and it’s one of the most useful things I’ve heard about coding with AI this year.
Lauren is a Principal Engineer at SpaceX. In October she shipped 3,127 PRs, and her team shipped almost 11,000. At the same time, the share of commits they had to revert fell from 0.98% to 0.47%.
Her point is that "slop" just means code you didn’t write and don’t trust. So the real job isn’t to review AI code harder. It’s to build a system where agents write good code by default and check their own work.
Her main rule: every time you correct an agent, don’t add another line to AGENTS.md.
Her team went from 7 custom lint rules to 167. They banned useEffect from product code and stopped parts of the app from importing each other. As she puts it, nuance tends to lead to slop.
My favorite story is about comments:
One agent left a quick setTimeout hack with a comment saying the real fix "would be a larger change". Five weeks later the app had 183 timers and lagged. So the team deleted almost every comment, from 65,824 down to 1,227.
The rest goes into verification: agents click through the real app, save screenshots as proof, and use formal methods where it matters. Lauren says most of her tokens now go to checking work, not writing it.
Her final warning is the part worth remembering: if your agents produce slop, automating them just gets you more slop, faster. Fix the foundation first, then scale.
Links:
Text version of the speech: Article
Full speech: Video
Posts with brief information: Post 1, Post 2
Grok Bot ( Pattern Repetition): Slop to Gold
ещё
39
1
1 17 ноя, 11:21
My X: https://x.com/0x_kaize
ещё
57
0
2 15 ноя, 06:21
Продолжить анализ этого канала
- История роста подписчиков
- Просмотры и охваты публикаций
- Кто упоминает канал и кого упоминает он
- Похожие каналы и подбор для рекламы
Открыть аналитику
Демо до 14 дней