Social Algorithm Updates: X, Reddit, TikTok, Pinterest, Facebook and LinkedIn

A day after publishing its ranking weights, X added a second dated update to the repo's README. The larger half of it is a country-specific change to who gets recommended.

From 16th August, Brazilian electoral law requires providers that run recommendation systems to exclude the official accounts of candidates registered with the Electoral Court, and those accounts' posts, from their recommendations. X says it is complying, and For You now runs a filter named Brazil2026ElectionFilter.

One word in the English README is worth reading carefully. It describes accounts "reported to Brazil's Electoral Court", which means declared to it in the candidate registry, not reported by users for misconduct. X's own Portuguese announcement is the clearer statement: these are the official accounts of registered candidates, and the measure is compliance with Resolution 23.610/2019 rather than a sanction or a change to X's own rules.

The code names every account. A hardcoded list of 665 user IDs, each annotated with its username, is checked against the author of every candidate post, and a post is dropped unless the viewer explicitly follows that account. The filter also catches reposts and quotes of a listed account, and replies whose conversation thread reaches one.

X is explicit that the accounts are not suspended and their content stays on their profiles, and equally explicit that the change will affect the visibility and reach of those profiles and their posts. The legal text quoted in the code carries one carve-out that marketers should note: the exclusion applies "except in cases of paid boosting".

The second half of the update answers something that happened to X within a day of the weights going public. Reading the published numbers as if a report cancels out a fixed number of likes spread quickly, and X has now written the rebuttal into the code itself.

The weights multiply predicted probabilities, not engagement tallies. As the new comment in the scoring code puts it, "the weights do not multiply raw engagement counts", so treating the ratio between two weights as a count equivalence is wrong.

The README is blunter about the specific claim, calling "1 report cancels out 468 likes" incorrect. It adds that the weights are "a multiple on your own predicted probability of Liking, Reporting, etc, which is substantially driven by your own behavior".

The comment also explains why the report weight is large in the first place: "the baseline probability of a Report is more than 1000x lower than a Like", so it has to be weighted up before the prediction can move the ranking at all.

That leads into the claim most likely to matter to anyone managing an account, that coordinated mass reporting or blocking can bury reach. X gives two reasons it does not work that way. Recommendations are personalised, so "reports from bad actors will primarily affect recommendations for users who are similar to the bad actors" rather than suppressing the post for everyone.

The second reason is narrower and more concrete. An action only counts in the recommendation system if it happens on a post that was served in the Home Timeline, so "directly navigating to a post (i.e., coordinating via groupchat) has no ranking impact", and a group cannot reliably make a target post appear in their own timelines to begin with.

X says it added the comments "so that LLMs or people reading it are more likely to understand it correctly", which is a fair description of what open-sourcing a ranking system now costs it. On the Brazil filter it makes the opposite point in its favour: "A benefit of open-source is that you can see that changes like this exist, and exactly how they work".

The repo's own README, dated 13th August 2026, lists three additions: the configuration parameters that blend predicted action values into a score, the code for systems that decide whether a post is filtered out of the For You feed, and real Phoenix training code in place of the old demonstration model, with synthetic data generation so a proof-of-concept training run can be done.

The new systems are the ones that decide visibility rather than order. Visibility filtering determines whether to show a post, drop it, or show it behind an interstitial, and the release also opens the machinery feeding it: rule engines that apply labels, models that score accounts, models that examine images and video, and an enforcement service.

The scoring detail is the part worth reading if you post on X. Phoenix predicts a probability for each action a viewer might take, across engagement, clicks, attention, following the author, and negative signals such as mute, block and report.

Those probabilities are summed using the now-published weights, positive for positive actions and negative for negative ones. Three adjustments follow: a decay applied to each post after an author's first, a discount for accounts the viewer does not follow, which also hits replies and reposts from accounts they do, and a boost for authors whose impressions sit below a threshold.

The filters are equally concrete. Posts older than 48 hours are dropped before scoring, along with posts the viewer has already been shown, their own posts, and reposts and replies from accounts they do not follow.

X's VP of Product Keith Coleman told TechCrunch that "You'll get the core ranking code that pulls posts and ranks them for any given user and assembles the feed", adding "This is the kind of thing that I think people will be fairly shocked that we are releasing". The code is Apache v2 licensed, and TechCrunch puts the expansion at roughly 10 to 15 times the previous size.

The paired transparency tool sits on an "Under the Hood" page in the app's settings. Accounts that posted 10 or more times in the past month can download aggregate statistics as a JSON file showing which visibility-affecting labels were applied to the account and its posts.

What is held back is stated openly, on the grounds that "people could use it to try to game the system": the Grox prompt files and some botmaker rules. Engadget adds that ads code and systems outside the For You timeline stay unreleased.

So this does not meet Musk's stated goal that "every line of code touching the X system will be open source and third-party audited" in August. X says third-party recommender-system experts pressure-tested the release beforehand, without saying how they were chosen or whether their findings will be published.

The paper, from eight Pinterest researchers, describes vision-language-model relevance judging deployed at Pinterest Search to evaluate A/B experiments: "VLMs can provide reliable relevance measurement for experiments while greatly improving the evaluation efficiency", with the approach cutting the minimum detectable effects in online experiments.

It extends the system Pinterest documented in December 2025, where fine-tuned open-source LLMs score query-and-Pin relevance on a five-level scale, matching human labels exactly 73.7% of the time and within one level 91.7% of the time.

The through-line for marketers: on Pinterest, "the algorithm" is unusually inspectable. The engineering blog has documented how search relevance is modeled, what feed ranking consumes (up to 16,000 lifelong actions per user) and now how quality itself is measured. The Pinterest section collects the trail.

Reddit's Q2 2026 shareholder letter reported its eighth consecutive quarter above 60% revenue growth: $804.9 million, net income of $253 million, daily uniques up 18% and logged-out daily uniques, the search-exposed cohort, up 27%.

The sentence that moved the market: "Search referrals were choppy in the quarter, and traffic was more volatile later in the quarter". Shares sank roughly 11% in after-hours trading despite above-consensus guidance.

Huffman's framing leaned into independence: "While our visibility in referral traffic remains low, we're not building for drive-by traffic. We're building a daily destination."

One omission is worth logging: this is the first letter in six quarters without a weekly-search-user or Reddit Answers figure, breaking the series that ran 60, 70, 75 and 80+ million.

The quarter's other Google story compounds the dependence question: a week earlier, per Wall Street Journal reporting carried by CNBC, Reddit had discussed cutting off Google's AI access to its content as the roughly $60-million-a-year data deal nears its end, with renewal talks under way and the stock sliding 8% on that report alone.

Tom Alison's post, "Connecting Real People on Facebook", announced a test that opens the app directly into full-screen video, with the promise that "Classic Feed will always be available as the second tab in the app, just one click away", plus an opt-out to remain Feed-first.

Rollout is staged: testing starts later this year in video-heavy international markets, with US exploration planned for the following year. The same post bundled Facebook Verified (a free selfie-verification badge), a Seller app for Marketplace and a Forum app for Groups.

Distribution-wise, the app's default entry surface becoming a ranked video feed is the biggest Facebook change since the December 2025 redesign moved search results to an immersive grid. Note the dating: much trade coverage describes an "August 2026 redesign", but Meta's own posts date the announcement to 9th December 2025 and this test to 24th July 2026, and no August feed post exists on the newsroom.