Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Cloudeflare грозит Google блокировкой трафика
Cloudflare объявил: с 15 сентября 2026 года ИИ-краулеры будут заблокированы по умолчанию на всех сайтах с рекламой — включая Googlebot, Applebot и Bingbot.
Главная претензия — к Google: один и тот же бот индексирует страницы и собирает данные для обучения нейросетей, что даёт поисковику нечестное преимущество.
Но есть нюанс, который меняет всю к…
➡️ Читайте на сайте: https://aff.top/blog/cloudeflare-grozit-google-blokirovkoi-trafika
🧠 Ещё больше инсайтов → в канале AFF.top
Cloudflare объявил: с 15 сентября 2026 года ИИ-краулеры будут заблокированы по умолчанию на всех сайтах с рекламой — включая Googlebot, Applebot и Bingbot.
Главная претензия — к Google: один и тот же бот индексирует страницы и собирает данные для обучения нейросетей, что даёт поисковику нечестное преимущество.
Но есть нюанс, который меняет всю к…
➡️ Читайте на сайте: https://aff.top/blog/cloudeflare-grozit-google-blokirovkoi-trafika
🧠 Ещё больше инсайтов → в канале AFF.top
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Гайд: как заработать первые деньги на Pornhub
Pornhub — самый посещаемый адалт-сайт в мире, и на нём действительно можно зарабатывать. Но схема устроена иначе, чем кажется.
Автор залил ролики, набрал 16 000 просмотров — и получил 47 центов встроенной монетизации. Реальные деньги были в другом.
Есть нюансы с верификацией, голосом в роликах и законодательством РФ, которые ломают большинство с…
➡️ Читайте на сайте: https://aff.top/blog/gaid-kak-zarabotat-pervye-dengi-na-pornhub
🧠 Ещё больше инсайтов → в канале AFF.top
Pornhub — самый посещаемый адалт-сайт в мире, и на нём действительно можно зарабатывать. Но схема устроена иначе, чем кажется.
Автор залил ролики, набрал 16 000 просмотров — и получил 47 центов встроенной монетизации. Реальные деньги были в другом.
Есть нюансы с верификацией, голосом в роликах и законодательством РФ, которые ломают большинство с…
➡️ Читайте на сайте: https://aff.top/blog/gaid-kak-zarabotat-pervye-dengi-na-pornhub
🧠 Ещё больше инсайтов → в канале AFF.top
E-E-A-T is not a ranking factor, and the people who said so were Google
The question: is there a number, a score, an 'E-E-A-T value' attached to your page somewhere in Google's index?
The documented answer is no. Google's own search liaisons have stated repeatedly that E-E-A-T is a concept, not a measurable signal. In Google's own words from the 'How Search Works' documentation, the systems use 'a variety of signals' that 'align with' the concept of trust — alignment, not implementation.
The supporting evidence is structural. The Quality Rater Guidelines explicitly tell raters their ratings 'do not directly impact' rankings (QRG overview section). Raters score sample results; those scores become training and evaluation data for engineers tuning systems. There is no path by which a rater's E-E-A-T judgment becomes a stored attribute on your URL.
Why practitioners get this wrong: the correlation is real even if the mechanism is indirect. Pages that raters would judge as high-E-E-A-T tend to also accumulate the link patterns, the brand searches, and the engagement that the actual systems reward. You can mistake the shadow for the object.
Caveat: 'concept not factor' is not license to ignore it. The concept describes the destination the systems are aimed at. Optimizing for the proxies the systems actually measure, while genuinely improving trust, is the only coherent strategy.
What we still don't know: the exact basket of measurable signals Google uses as proxies for trust. The 2024 leak named thousands of features but did not hand us the weighting, and weighting is where the truth lives.
The question: is there a number, a score, an 'E-E-A-T value' attached to your page somewhere in Google's index?
The documented answer is no. Google's own search liaisons have stated repeatedly that E-E-A-T is a concept, not a measurable signal. In Google's own words from the 'How Search Works' documentation, the systems use 'a variety of signals' that 'align with' the concept of trust — alignment, not implementation.
The supporting evidence is structural. The Quality Rater Guidelines explicitly tell raters their ratings 'do not directly impact' rankings (QRG overview section). Raters score sample results; those scores become training and evaluation data for engineers tuning systems. There is no path by which a rater's E-E-A-T judgment becomes a stored attribute on your URL.
Why practitioners get this wrong: the correlation is real even if the mechanism is indirect. Pages that raters would judge as high-E-E-A-T tend to also accumulate the link patterns, the brand searches, and the engagement that the actual systems reward. You can mistake the shadow for the object.
Caveat: 'concept not factor' is not license to ignore it. The concept describes the destination the systems are aimed at. Optimizing for the proxies the systems actually measure, while genuinely improving trust, is the only coherent strategy.
What we still don't know: the exact basket of measurable signals Google uses as proxies for trust. The 2024 leak named thousands of features but did not hand us the weighting, and weighting is where the truth lives.
The asymmetry nobody plans for: reputation can only hurt you faster than it helps
The question: in the Quality Rater Guidelines, what triggers the harshest rating — and is the bar symmetric?
It is not symmetric. The QRG (section 4 on Lowest quality) instructs raters to assign Lowest when a page or its creator has a sufficiently negative reputation, and crucially, independent negative reputation can override otherwise good content. But the reverse is weaker: a strong positive reputation supports a High rating, it rarely forces one on its own. Trust is treated as a veto, not a multiplier.
The evidence for this design: the guidelines spend far more pages on detecting harm, deception, and untrustworthiness than on rewarding excellence. Raters are told to actively search outside the site — reviews, news, references — specifically looking for evidence the creator is untrustworthy. There is no equivalent mandate to hunt for proof of greatness.
Why this matters for strategy: it means reputation defense has higher expected value than reputation building past a threshold. One credible expose, one pattern of scam reports, can floor an entire domain's rated quality regardless of on-page polish.
Counter-evidence: this is rater behavior, not confirmed algorithmic behavior. Whether production systems implement a true reputation veto at scale is unproven — they would need a reliable entity-reputation source, which is hard.
Caveat: 'reputation' in the QRG is specifically independent, off-site reputation. Self-published testimonials do not count and are explicitly discounted.
What we still don't know: how systems approximate off-site reputation at web scale without a curated knowledge source for most entities.
The question: in the Quality Rater Guidelines, what triggers the harshest rating — and is the bar symmetric?
It is not symmetric. The QRG (section 4 on Lowest quality) instructs raters to assign Lowest when a page or its creator has a sufficiently negative reputation, and crucially, independent negative reputation can override otherwise good content. But the reverse is weaker: a strong positive reputation supports a High rating, it rarely forces one on its own. Trust is treated as a veto, not a multiplier.
The evidence for this design: the guidelines spend far more pages on detecting harm, deception, and untrustworthiness than on rewarding excellence. Raters are told to actively search outside the site — reviews, news, references — specifically looking for evidence the creator is untrustworthy. There is no equivalent mandate to hunt for proof of greatness.
Why this matters for strategy: it means reputation defense has higher expected value than reputation building past a threshold. One credible expose, one pattern of scam reports, can floor an entire domain's rated quality regardless of on-page polish.
Counter-evidence: this is rater behavior, not confirmed algorithmic behavior. Whether production systems implement a true reputation veto at scale is unproven — they would need a reliable entity-reputation source, which is hard.
Caveat: 'reputation' in the QRG is specifically independent, off-site reputation. Self-published testimonials do not count and are explicitly discounted.
What we still don't know: how systems approximate off-site reputation at web scale without a curated knowledge source for most entities.
YMYL is a spectrum with a multiplier, not a category you're in or out of
The question: is a page either YMYL or not, like a checkbox?
The guidelines describe it as a gradient. Per the QRG's treatment of Your Money or Your Life topics, a page can be 'clearly YMYL,' 'possibly YMYL,' or 'clearly not YMYL,' and raters are told to consider the potential to cause harm — to the individual, to other people, or to society — as a matter of degree. A page on choosing running shoes is low on the spectrum; a page on medication interactions is at the ceiling.
The operational consequence is a multiplier effect. The standard of trust required does not switch on; it scales with potential harm. The same thin author bio that is acceptable on a hobby-craft page becomes disqualifying on a page about cancer treatment, because the cost of being wrong differs by orders of magnitude.
Supporting evidence: the four harm types in the guidelines — health/safety, financial, civic/societal, and 'other' high-stakes — give raters a structured way to place a topic on the gradient rather than sorting into two bins.
Counter-evidence and caveat: this is the rater's framework. We have no public confirmation that systems compute a continuous YMYL harm-score per query. They may use topical classifiers that behave more like coarse buckets.
What we still don't know: where the boundaries sit for ambiguous topics — pet health, amateur legal advice, nutrition for general wellness. These sit in the murky middle of the spectrum and almost certainly get inconsistent treatment.
The question: is a page either YMYL or not, like a checkbox?
The guidelines describe it as a gradient. Per the QRG's treatment of Your Money or Your Life topics, a page can be 'clearly YMYL,' 'possibly YMYL,' or 'clearly not YMYL,' and raters are told to consider the potential to cause harm — to the individual, to other people, or to society — as a matter of degree. A page on choosing running shoes is low on the spectrum; a page on medication interactions is at the ceiling.
The operational consequence is a multiplier effect. The standard of trust required does not switch on; it scales with potential harm. The same thin author bio that is acceptable on a hobby-craft page becomes disqualifying on a page about cancer treatment, because the cost of being wrong differs by orders of magnitude.
Supporting evidence: the four harm types in the guidelines — health/safety, financial, civic/societal, and 'other' high-stakes — give raters a structured way to place a topic on the gradient rather than sorting into two bins.
Counter-evidence and caveat: this is the rater's framework. We have no public confirmation that systems compute a continuous YMYL harm-score per query. They may use topical classifiers that behave more like coarse buckets.
What we still don't know: where the boundaries sit for ambiguous topics — pet health, amateur legal advice, nutrition for general wellness. These sit in the murky middle of the spectrum and almost certainly get inconsistent treatment.
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Сбер запустит свой криптокошелёк
Сбер готов запустить криптокошелёк — инфраструктура уже есть. Ждут только закона о регулировании крипты, который планируют принять к 1 сентября 2026 года.
Хранить и, судя по всему, обменивать крипту можно будет прямо в приложении — без сторонних обменников.
Но есть один нюанс, из-за которого обменники никуда не денутся. 🔍
➡️ Читайте на сайте: https://aff.top/blog/sber-zapustit-svoi-kriptokoshelek
🧠 Ещё больше инсайтов → в канале AFF.top
Сбер готов запустить криптокошелёк — инфраструктура уже есть. Ждут только закона о регулировании крипты, который планируют принять к 1 сентября 2026 года.
Хранить и, судя по всему, обменивать крипту можно будет прямо в приложении — без сторонних обменников.
Но есть один нюанс, из-за которого обменники никуда не денутся. 🔍
➡️ Читайте на сайте: https://aff.top/blog/sber-zapustit-svoi-kriptokoshelek
🧠 Ещё больше инсайтов → в канале AFF.top
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Индия потребовала от Telegram удалять пиратский контент
Индия потребовала от Telegram удалять пиратский контент — претензия в том, что платформа не ограничивает размер файлов, что позволяет свободно распространять фильмы.
Дуров ответил, что Telegram годами работает в Индии без какой-либо коммерческой выгоды для себя.
Почему давление началось именно сейчас — вопрос открытый. Возможный ответ — в блоге.
➡️ Читайте на сайте: https://aff.top/blog/indiia-potrebovala-ot-telegram-udaliat-piratskii-kontent
🧠 Ещё больше инсайтов → в канале AFF.top
Индия потребовала от Telegram удалять пиратский контент — претензия в том, что платформа не ограничивает размер файлов, что позволяет свободно распространять фильмы.
Дуров ответил, что Telegram годами работает в Индии без какой-либо коммерческой выгоды для себя.
Почему давление началось именно сейчас — вопрос открытый. Возможный ответ — в блоге.
➡️ Читайте на сайте: https://aff.top/blog/indiia-potrebovala-ot-telegram-udaliat-piratskii-kontent
🧠 Ещё больше инсайтов → в канале AFF.top
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
Google ads меняет стратегию по конверсиям
Google меняет логику автоматических стратегий ставок: с 17 августа 2026 года кампании будут строже придерживаться указанного целевого CPA, а не давать лиды по минимально возможной цене.
Если сейчас твоя кампания даёт лиды по $5, а цель стоит $10 — после обновления алгоритм «поднимет» фактическую стоимость лида к целевой, зато отдаст больше трафик…
➡️ Читайте на сайте: https://aff.top/blog/google-ads-meniaet-strategiiu-po-konversiiam
🧠 Ещё больше инсайтов → в канале AFF.top
Google меняет логику автоматических стратегий ставок: с 17 августа 2026 года кампании будут строже придерживаться указанного целевого CPA, а не давать лиды по минимально возможной цене.
Если сейчас твоя кампания даёт лиды по $5, а цель стоит $10 — после обновления алгоритм «поднимет» фактическую стоимость лида к целевой, зато отдаст больше трафик…
➡️ Читайте на сайте: https://aff.top/blog/google-ads-meniaet-strategiiu-po-konversiiam
🧠 Ещё больше инсайтов → в канале AFF.top
A byline is a string. An author is an entity. Search cares about the second one.
The question: does adding an author name to a page do anything, or does it require the name to resolve to a known entity?
The distinction is the whole game. A byline is text on a page. An entity is a node in a knowledge graph with relationships — publications, affiliations, a disambiguated identity. Google's systems, per the entity-recognition work described across its patents and the Knowledge Graph documentation, operate on entities, not strings. An author who exists only as words on your own site has no independent corroboration to draw on.
The evidence: the QRG instructs raters to research the reputation of the content creator using independent sources. A bare byline gives them nothing to research. An author who is an established entity — with bylines across reputable sites, a Wikipedia or Wikidata node, conference talks indexed and findable — provides corroborating evidence that raters and systems can both reference.
Counter-evidence worth stating: most ranking content has no notable author entity and ranks fine. Entity-level authorship matters most precisely where YMYL stakes are high; elsewhere its marginal value is small and easily overstated.
Caveat: building an author entity is slow and cannot be faked with markup alone. sameAs links to profiles you control are weak; what counts is independent web consensus that the entity exists and is what it claims.
What we still don't know: how heavily, if at all, production ranking weights author-entity corroboration versus site-level signals. The honest answer is that site-level trust still appears to dominate for most queries.
The question: does adding an author name to a page do anything, or does it require the name to resolve to a known entity?
The distinction is the whole game. A byline is text on a page. An entity is a node in a knowledge graph with relationships — publications, affiliations, a disambiguated identity. Google's systems, per the entity-recognition work described across its patents and the Knowledge Graph documentation, operate on entities, not strings. An author who exists only as words on your own site has no independent corroboration to draw on.
The evidence: the QRG instructs raters to research the reputation of the content creator using independent sources. A bare byline gives them nothing to research. An author who is an established entity — with bylines across reputable sites, a Wikipedia or Wikidata node, conference talks indexed and findable — provides corroborating evidence that raters and systems can both reference.
Counter-evidence worth stating: most ranking content has no notable author entity and ranks fine. Entity-level authorship matters most precisely where YMYL stakes are high; elsewhere its marginal value is small and easily overstated.
Caveat: building an author entity is slow and cannot be faked with markup alone. sameAs links to profiles you control are weak; what counts is independent web consensus that the entity exists and is what it claims.
What we still don't know: how heavily, if at all, production ranking weights author-entity corroboration versus site-level signals. The honest answer is that site-level trust still appears to dominate for most queries.
Forwarded from AFF.TOP
This media is not supported in your browser
VIEW IN TELEGRAM
В Codex внедрят GPT-5.6 Ultra
OpenAI добавит в Codex эксклюзивную версию GPT-5.6 Sol Ultra — не ту, что выйдет в паблик, а отдельную, усиленную модель.
Два ключевых режима: расширенные рассуждения (модель думает дольше) и мульти-агентная работа с параллельными субагентами. Релиз ожидается 7–9 июля 2026.
Но есть один нюанс, который OpenAI пока не раскрывает 👀 Подробности — в …
➡️ Читайте на сайте: https://aff.top/blog/v-codex-vnedriat-gpt-5-6-ultra
🧠 Ещё больше инсайтов → в канале AFF.top
OpenAI добавит в Codex эксклюзивную версию GPT-5.6 Sol Ultra — не ту, что выйдет в паблик, а отдельную, усиленную модель.
Два ключевых режима: расширенные рассуждения (модель думает дольше) и мульти-агентная работа с параллельными субагентами. Релиз ожидается 7–9 июля 2026.
Но есть один нюанс, который OpenAI пока не раскрывает 👀 Подробности — в …
➡️ Читайте на сайте: https://aff.top/blog/v-codex-vnedriat-gpt-5-6-ultra
🧠 Ещё больше инсайтов → в канале AFF.top
Can a machine tell genuine first-hand experience from a confident fabrication?
The question: Google says it values 'experience' — but can any system actually detect it in text, or only proxy for it?
The honest position is: only proxy, weakly. There is no known classifier that reads a paragraph and outputs 'this person truly did this.' What systems can do is detect the textual signatures that correlate with experience — specific sensory detail, original photographs with consistent metadata, mention of edge cases a non-practitioner wouldn't anticipate, and the absence of the hedged generality typical of synthesized content.
The evidence that these are proxies, not the thing itself: every one of those signatures can be manufactured. A skilled writer or a capable model can produce specific detail and plausible edge cases without ever touching the subject. This is the central epistemological problem of the experience signal.
Where it still bites: original media is the hardest signature to fake at scale and cheapest to verify. A review with twelve original photos of a product in real environments, with consistent lighting and incidental background detail, is expensive to fabricate and easy for both raters and reverse-image systems to corroborate.
Caveat: correlation is not causation here. Pages with original media rank better, but original media also correlates with overall production investment, which correlates with many other quality proxies. Isolating the experience effect is not something anyone has cleanly done in public.
What we still don't know: whether the rise of generative media erodes original-media as a trust signal faster than detection improves. This is an active arms race with no settled outcome.
The question: Google says it values 'experience' — but can any system actually detect it in text, or only proxy for it?
The honest position is: only proxy, weakly. There is no known classifier that reads a paragraph and outputs 'this person truly did this.' What systems can do is detect the textual signatures that correlate with experience — specific sensory detail, original photographs with consistent metadata, mention of edge cases a non-practitioner wouldn't anticipate, and the absence of the hedged generality typical of synthesized content.
The evidence that these are proxies, not the thing itself: every one of those signatures can be manufactured. A skilled writer or a capable model can produce specific detail and plausible edge cases without ever touching the subject. This is the central epistemological problem of the experience signal.
Where it still bites: original media is the hardest signature to fake at scale and cheapest to verify. A review with twelve original photos of a product in real environments, with consistent lighting and incidental background detail, is expensive to fabricate and easy for both raters and reverse-image systems to corroborate.
Caveat: correlation is not causation here. Pages with original media rank better, but original media also correlates with overall production investment, which correlates with many other quality proxies. Isolating the experience effect is not something anyone has cleanly done in public.
What we still don't know: whether the rise of generative media erodes original-media as a trust signal faster than detection improves. This is an active arms race with no settled outcome.