Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Компания Meta выпустила Muse Spark 1.1
Meta выпустила Muse Spark 1.1 почти одновременно с новой ChatGPT-5.6. Это мультимодальный агент, который сам дробит задачу на подзадачи и распределяет их между субагентами.
Стоимость тоже заметно ниже топовых западных моделей: $1.25 за миллион входных токенов и $4.25 за миллион выходных.
Но главный вопрос — насколько она реально сильна на фоне к…
➡️ Читайте на сайте: https://aff.top/blog/kompaniia-meta-vypustila-muse-spark-1-1
🧠 Ещё больше инсайтов → в канале AFF.top
Meta выпустила Muse Spark 1.1 почти одновременно с новой ChatGPT-5.6. Это мультимодальный агент, который сам дробит задачу на подзадачи и распределяет их между субагентами.
Стоимость тоже заметно ниже топовых западных моделей: $1.25 за миллион входных токенов и $4.25 за миллион выходных.
Но главный вопрос — насколько она реально сильна на фоне к…
➡️ Читайте на сайте: https://aff.top/blog/kompaniia-meta-vypustila-muse-spark-1-1
🧠 Ещё больше инсайтов → в канале AFF.top
Every join cohort has a half-life
Servers report headline member counts. The cohort lens tells a more honest, more useful story.
What the data shows
When joins are bucketed by month and tracked, most communities show a brutal early cliff: a large share of any cohort goes inactive within the first 7 days, and the curve then flattens into a slow-decaying tail. A useful summary statistic is the cohort half-life — the time for a join-month group to lose half its actives. Healthy topic servers often land somewhere in the multi-month range; servers riding a viral spike can have half-lives measured in days.
Why it happens
The early cliff is filtration — wrong-fit joiners leave fast. The tail is genuine attrition of fit members and is the number worth defending. Spike-driven growth inflates the top of the funnel with low-intent joiners, which is why raw member count and viral moments flatter a server that's actually bleeding.
The caveat
"Inactive" is a definitional choice (no message? no visit? no reaction?), and different thresholds move the curve a lot. Without server-side analytics, most owners can only approximate this.
Open question: is the early cliff worth fighting at all — or is it healthy filtration that you should optimize the tail around instead?
Servers report headline member counts. The cohort lens tells a more honest, more useful story.
What the data shows
When joins are bucketed by month and tracked, most communities show a brutal early cliff: a large share of any cohort goes inactive within the first 7 days, and the curve then flattens into a slow-decaying tail. A useful summary statistic is the cohort half-life — the time for a join-month group to lose half its actives. Healthy topic servers often land somewhere in the multi-month range; servers riding a viral spike can have half-lives measured in days.
Why it happens
The early cliff is filtration — wrong-fit joiners leave fast. The tail is genuine attrition of fit members and is the number worth defending. Spike-driven growth inflates the top of the funnel with low-intent joiners, which is why raw member count and viral moments flatter a server that's actually bleeding.
The caveat
"Inactive" is a definitional choice (no message? no visit? no reaction?), and different thresholds move the curve a lot. Without server-side analytics, most owners can only approximate this.
Open question: is the early cliff worth fighting at all — or is it healthy filtration that you should optimize the tail around instead?
The verification gate's two-sided cost
Captcha and verification gates are pitched as pure anti-spam wins. The funnel data shows they're a trade, not a free lunch.
What the data shows
Servers that added a hard verification step (react-to-enter, captcha bot, rules-gate) report large drops in spam and raid damage — that part is real and well-replicated. But join-completion telemetry shows meaningful fall-off at the gate: a non-trivial share of legitimate joiners abandon before completing verification, with the loss heaviest on mobile and for multi-step flows.
Why it happens
Every added step in any funnel sheds users — this is the oldest finding in conversion research, and a Discord gate is just another funnel step. The gate filters bots and low-intent humans, which is partly the point — but "low-intent" overlaps with "shy newcomer who'd have warmed up."
The caveat
The abandoned joiners may be near-worthless (bots, drive-bys), in which case the loss is illusory. We rarely know the counterfactual value of who we filtered. Telegram's join-request + captcha bots face the identical trade-off, under-studied at scale.
Open question: what's the lightest gate that stops raids — a single one-tap reaction versus a full captcha — and where exactly does the spam-blocked curve cross the humans-lost curve?
Captcha and verification gates are pitched as pure anti-spam wins. The funnel data shows they're a trade, not a free lunch.
What the data shows
Servers that added a hard verification step (react-to-enter, captcha bot, rules-gate) report large drops in spam and raid damage — that part is real and well-replicated. But join-completion telemetry shows meaningful fall-off at the gate: a non-trivial share of legitimate joiners abandon before completing verification, with the loss heaviest on mobile and for multi-step flows.
Why it happens
Every added step in any funnel sheds users — this is the oldest finding in conversion research, and a Discord gate is just another funnel step. The gate filters bots and low-intent humans, which is partly the point — but "low-intent" overlaps with "shy newcomer who'd have warmed up."
The caveat
The abandoned joiners may be near-worthless (bots, drive-bys), in which case the loss is illusory. We rarely know the counterfactual value of who we filtered. Telegram's join-request + captcha bots face the identical trade-off, under-studied at scale.
Open question: what's the lightest gate that stops raids — a single one-tap reaction versus a full captcha — and where exactly does the spam-blocked curve cross the humans-lost curve?
Servers have a heartbeat, and it's geographic
Activity isn't uniform across the day, and the shape of the curve carries more diagnostic information than the average.
What the data shows
Hourly message histograms across servers reveal that "global" communities frequently have a single dominant timezone cluster hiding under the label — one peak, one trough, and long dead hours. Servers that felt alive 18+ hours a day almost always had at least two distinct geographic clusters producing a double-peaked curve. Single-peak servers had longer dead zones where conversation died and didn't restart.
Why it happens
Conversation needs overlapping presence. A single timezone cluster guarantees a daily window where no one's around to sustain a thread, and a thread that dies overnight rarely resurrects. Two clusters hand the baton across the clock. Telegram's async, feed-like consumption tolerates dead hours far better — a post at 3am still gets seen at 9am — which is a structural advantage for genuinely global audiences.
The caveat
Locale data is inferred and noisy; VPNs and night-owls blur it. And a double peak could just be a proxy for "bigger," which brings its own activity.
Open question: should a single-timezone server deliberately recruit a second cluster — or specialize and accept the dead hours rather than dilute its core?
Activity isn't uniform across the day, and the shape of the curve carries more diagnostic information than the average.
What the data shows
Hourly message histograms across servers reveal that "global" communities frequently have a single dominant timezone cluster hiding under the label — one peak, one trough, and long dead hours. Servers that felt alive 18+ hours a day almost always had at least two distinct geographic clusters producing a double-peaked curve. Single-peak servers had longer dead zones where conversation died and didn't restart.
Why it happens
Conversation needs overlapping presence. A single timezone cluster guarantees a daily window where no one's around to sustain a thread, and a thread that dies overnight rarely resurrects. Two clusters hand the baton across the clock. Telegram's async, feed-like consumption tolerates dead hours far better — a post at 3am still gets seen at 9am — which is a structural advantage for genuinely global audiences.
The caveat
Locale data is inferred and noisy; VPNs and night-owls blur it. And a double peak could just be a proxy for "bigger," which brings its own activity.
Open question: should a single-timezone server deliberately recruit a second cluster — or specialize and accept the dead hours rather than dilute its core?
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Российские букмекеры увеличили закуп трафика с мобильных приложений
Российские букмекеры в 1 квартале 2026 года заметно нарастили закупку трафика из мобильных приложений. На фоне ужесточения регулирования они смещают бюджеты в новые каналы, где ещё есть живой трафик.
По данным UMG, доля in-app-рекламы выросла с 3-4% до 5-6% при объёме рынка около 10 млрд рублей. Но это может быть только начало — в блоге разбираем…
➡️ Читайте на сайте: https://aff.top/blog/rossiiskie-bukmekery-uvelichili-zakup-trafika-s-mobilnykh-prilozhenii
🧠 Ещё больше инсайтов → в канале AFF.top
Российские букмекеры в 1 квартале 2026 года заметно нарастили закупку трафика из мобильных приложений. На фоне ужесточения регулирования они смещают бюджеты в новые каналы, где ещё есть живой трафик.
По данным UMG, доля in-app-рекламы выросла с 3-4% до 5-6% при объёме рынка около 10 млрд рублей. Но это может быть только начало — в блоге разбираем…
➡️ Читайте на сайте: https://aff.top/blog/rossiiskie-bukmekery-uvelichili-zakup-trafika-s-mobilnykh-prilozhenii
🧠 Ещё больше инсайтов → в канале AFF.top
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Mia Khalifa стала амбпссадором 1win
Mia Khalifa снова засветилась рядом с 1win: в Instagram она показала цепочку с логотипом бренда и статус «VIP 1win».
Параллельно всплыла история с ее ставкой на Испанию на ЧМ — по данным поста, выигрыш мог составить 200к$.
Официального подтверждения амбассадорства пока нет, но для арбитражников это уже повод для новых креативов. Что именно здесь…
➡️ Читайте на сайте: https://aff.top/blog/mia-khalifa-stala-ambpssadorom-1win
🧠 Ещё больше инсайтов → в канале AFF.top
Mia Khalifa снова засветилась рядом с 1win: в Instagram она показала цепочку с логотипом бренда и статус «VIP 1win».
Параллельно всплыла история с ее ставкой на Испанию на ЧМ — по данным поста, выигрыш мог составить 200к$.
Официального подтверждения амбассадорства пока нет, но для арбитражников это уже повод для новых креативов. Что именно здесь…
➡️ Читайте на сайте: https://aff.top/blog/mia-khalifa-stala-ambpssadorom-1win
🧠 Ещё больше инсайтов → в канале AFF.top
Forwarded from КРАВЧЕНКО
Дорогие коллеги и партнеры,
Наш маршрут конференций за последние недели, получился особенно насыщенным.
Со стендами PoshFriends мы побывали на MAC и GGate, а затем продолжили встречи уже в полях iGB Live в Лондоне.
В Ереване увиделись с любимыми SEO-командами, попробовали местные вина, обменялись новостями и зарядились энергией УБТ-команд.
В Тбилиси обсуждали тренды, новые связки и совместные планы, встречались с действующими партнерами и знакомились с новыми. А за настроение на стенде отвечала Черемша, которая чуть не стала маскотом одного из наших продуктов. С этой задачей, кажется, справилась лучше всех.
В Лондоне все было уже по-деловому. Провели серию встреч с топ-партнерами, обсудили Японию, бурж и новые точки роста. География интересов растет, планы становятся амбициознее. Воротники, как выяснилось, нагладили не зря.
Спасибо всем, с кем удалось увидеться на этом маршруте. За открытые разговоры, новые идеи, доверие и планы, которые постепенно превращаются в реальные проекты.
Конференционный сезон продолжается. Скоро увидимся снова.
Всегда ваши, Команда Posh Friends 🤝
Наш маршрут конференций за последние недели, получился особенно насыщенным.
Со стендами PoshFriends мы побывали на MAC и GGate, а затем продолжили встречи уже в полях iGB Live в Лондоне.
В Ереване увиделись с любимыми SEO-командами, попробовали местные вина, обменялись новостями и зарядились энергией УБТ-команд.
В Тбилиси обсуждали тренды, новые связки и совместные планы, встречались с действующими партнерами и знакомились с новыми. А за настроение на стенде отвечала Черемша, которая чуть не стала маскотом одного из наших продуктов. С этой задачей, кажется, справилась лучше всех.
В Лондоне все было уже по-деловому. Провели серию встреч с топ-партнерами, обсудили Японию, бурж и новые точки роста. География интересов растет, планы становятся амбициознее. Воротники, как выяснилось, нагладили не зря.
Спасибо всем, с кем удалось увидеться на этом маршруте. За открытые разговоры, новые идеи, доверие и планы, которые постепенно превращаются в реальные проекты.
Конференционный сезон продолжается. Скоро увидимся снова.
Всегда ваши, Команда Posh Friends 🤝
Reactions are a poll you didn't know you were running
Emoji reactions get dismissed as fluff. As a measurement instrument, they're underrated — and easy to misread.
What the data shows
Reaction-to-message ratios track engagement quality more sensitively than message count in several studied servers. A rising ratio often precedes a healthy stretch; a message volume that holds steady while reactions fall has shown up as an early warning before activity slumps — the room is still talking but no longer listening to each other.
Why it happens
A reaction is the cheapest possible signal of attention — near-zero cost, so it captures the marginal lurker who'd never type. When even that signal dries up, it suggests the silent majority has disengaged before the vocal minority notices. Telegram channels expose this even more starkly: with reactions and view counts but few replies, the reaction/view ratio is sometimes the only engagement signal available.
The caveat
Reaction culture is server-specific — some communities barely react regardless of health, so absolute thresholds don't transfer. The signal is in the trend, not the level, and small servers are too noisy to read week-to-week.
Open question: is a falling reaction/message ratio a true leading indicator of slump — or does it just track a normal seasonal dip that would self-correct?
Emoji reactions get dismissed as fluff. As a measurement instrument, they're underrated — and easy to misread.
What the data shows
Reaction-to-message ratios track engagement quality more sensitively than message count in several studied servers. A rising ratio often precedes a healthy stretch; a message volume that holds steady while reactions fall has shown up as an early warning before activity slumps — the room is still talking but no longer listening to each other.
Why it happens
A reaction is the cheapest possible signal of attention — near-zero cost, so it captures the marginal lurker who'd never type. When even that signal dries up, it suggests the silent majority has disengaged before the vocal minority notices. Telegram channels expose this even more starkly: with reactions and view counts but few replies, the reaction/view ratio is sometimes the only engagement signal available.
The caveat
Reaction culture is server-specific — some communities barely react regardless of health, so absolute thresholds don't transfer. The signal is in the trend, not the level, and small servers are too noisy to read week-to-week.
Open question: is a falling reaction/message ratio a true leading indicator of slump — or does it just track a normal seasonal dip that would self-correct?
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Короткий домен Telegram перестал работать
Telegram лишился домена t.me: он разделегирован и больше не работает на уровне регистратора. Платформа срочно переезжает на telegram.me, а владельцам крупных каналов стоит обновить публичные ссылки. Сроки восстановления неизвестны, и есть риск, что t.me не вернётся вовсе на фоне давления на Telegram.
➡️ Читайте на сайте: https://aff.top/blog/korotkii-domen-telegram-perestal-rabotat
🧠 Ещё больше инсайтов → в канале AFF.top
Telegram лишился домена t.me: он разделегирован и больше не работает на уровне регистратора. Платформа срочно переезжает на telegram.me, а владельцам крупных каналов стоит обновить публичные ссылки. Сроки восстановления неизвестны, и есть риск, что t.me не вернётся вовсе на фоне давления на Telegram.
➡️ Читайте на сайте: https://aff.top/blog/korotkii-domen-telegram-perestal-rabotat
🧠 Ещё больше инсайтов → в канале AFF.top
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Youtube тестирует поиск с AI
YouTube начал тестировать Ask YouTube — поиск с ИИ, где можно задавать вопросы обычным языком и получать не список ссылок, а готовую подборку видео и фрагментов.
Фича уже доступна в США и работает на сложные запросы: если нужно, ИИ уточняет вопрос и подсказывает следующий шаг.
Что это значит для поиска на YouTube и когда новинка дойдёт до других…
➡️ Читайте на сайте: https://aff.top/blog/youtube-testiruet-poisk-s-ai
🧠 Ещё больше инсайтов → в канале AFF.top
YouTube начал тестировать Ask YouTube — поиск с ИИ, где можно задавать вопросы обычным языком и получать не список ссылок, а готовую подборку видео и фрагментов.
Фича уже доступна в США и работает на сложные запросы: если нужно, ИИ уточняет вопрос и подсказывает следующий шаг.
Что это значит для поиска на YouTube и когда новинка дойдёт до других…
➡️ Читайте на сайте: https://aff.top/blog/youtube-testiruet-poisk-s-ai
🧠 Ещё больше инсайтов → в канале AFF.top
Threads remember; channels forget
Discord shipped threads as a clutter fix. Their more interesting effect is on a community's collective memory.
What the data shows
Servers that moved long-running topics into threads showed two changes: the main channel stayed readable, and — less obviously — old conversations stayed resumable. Thread-organized knowledge got revisited and replied-to days or weeks later at higher rates than the same content left to scroll past in a flat channel.
Why it happens
A flat channel has no addressable memory — once a topic scrolls off, it's effectively gone, because there's nowhere to "return to" it. A thread is a persistent container with its own notification context, so a question asked Tuesday can still get answered Friday without being lost. This is the inverse of Telegram's weakness: Telegram's reply chains exist but don't aggregate into browsable, named containers, so knowledge there decays into the timeline.
The caveat
Threads fragment discovery — a great answer buried in a thread is invisible to someone who never opens it, trading scroll-loss for click-loss. The net effect on findability is genuinely unsettled, and over-threading recreates the dead-channel problem at smaller scale.
Open question: is the right structure few-channels-many-threads, or does thread sprawl just relocate the silence problem one level down?
Discord shipped threads as a clutter fix. Their more interesting effect is on a community's collective memory.
What the data shows
Servers that moved long-running topics into threads showed two changes: the main channel stayed readable, and — less obviously — old conversations stayed resumable. Thread-organized knowledge got revisited and replied-to days or weeks later at higher rates than the same content left to scroll past in a flat channel.
Why it happens
A flat channel has no addressable memory — once a topic scrolls off, it's effectively gone, because there's nowhere to "return to" it. A thread is a persistent container with its own notification context, so a question asked Tuesday can still get answered Friday without being lost. This is the inverse of Telegram's weakness: Telegram's reply chains exist but don't aggregate into browsable, named containers, so knowledge there decays into the timeline.
The caveat
Threads fragment discovery — a great answer buried in a thread is invisible to someone who never opens it, trading scroll-loss for click-loss. The net effect on findability is genuinely unsettled, and over-threading recreates the dead-channel problem at smaller scale.
Open question: is the right structure few-channels-many-threads, or does thread sprawl just relocate the silence problem one level down?