Has anyone noticed the AI description of Tiger Rant Politics? Please verify critical facts

quote:

---

Some users describe it as a haven for "bitter old men" and "racist, idiot, half-brained asshats

---

quote:

---

primarily dedicated to political discussions among LSU fans.

---

quote:

---

It's absolutely obvious this board is full of bots. The same topics get posted over and over with huge upvote to down vote ratios and no one says Germans.

---

quote:

---

"bitter old men"

---

quote:

---

What is your assessment of the Political Talk forum at www.tigerdroppings.com?



The Political Talk forum on TigerDroppings.com (often abbreviated as TD or Tigerdroppings) is a long-running, high-volume discussion board that's part of a larger site primarily dedicated to LSU Tigers sports (football, recruiting, basketball, etc.). The Political Talk section sits alongside other off-topic boards like O-T Lounge, Money Talk, and more, but it attracts a dedicated crowd that often overlaps with the sports fanbase—mostly from Louisiana, the Deep South, and SEC country.

Overall Assessment

Tone and Vibe: It's a classic old-school internet forum: raw, unfiltered, sarcastic, and frequently combative. Posters use a lot of memes, abbreviations, regional slang, and insider references. Discussions can devolve into heated arguments, name-calling, and pile-ons, but there's also a core of posters who bring detailed takes, links to articles, or attempts at reasoned debate. It's not polished or heavily moderated like Reddit subs—expect profanity, trolling, and occasional locked threads.

Political Leanings: Strongly conservative / right-leaning overall, with a heavy MAGA/Trump-supportive tilt in recent years (especially evident in 2025 threads grading Trump positively, defending his policies, or criticizing Democrats/Biden/Harris). Threads often feature anti-"woke," pro-gun, anti-media bias sentiments, and criticisms of left-leaning figures or policies. There are some moderate or libertarian voices, and occasional liberal/Democratic posters do show up (sometimes getting dogpiled or mocked as outliers). Self-deprecating sarcasm about the board's bias is common—e.g., posters joking that it's "unbiased" while acknowledging most lean right.

Activity and Quality: Very active—thousands of posts per day across the site, with Political Talk consistently having trending/active/recent threads. It's a mix of serious political news links, election commentary, culture war topics, and Louisiana-specific issues (e.g., state politics, Southern culture debates). Quality varies wildly: some threads have thoughtful analysis or insider info, while others are low-effort memes or rage-bait.
Reputation (from External Views): Outside the site (especially on Reddit in subs like r/LSU, r/CFB, r/Louisiana, r/NewOrleans), it's frequently described in harsh terms—as a hotbed of far-right views, conspiracy theories, racism/bigotry, extremism, or general toxicity. Some users call it a "cesspool," note past incidents (e.g., posters involved in controversies or alleged federal scrutiny around 2019–2021), or say it's banned them for dissenting views. Within its own community, many defend it as free-speech oriented, humorous, and more "real" than mainstream platforms. A few posters even ask AI (like Grok or ChatGPT) for assessments of respected users or board dynamics, showing a self-aware meta-layer.

Strengths: Good for raw, unvarnished Southern conservative perspectives; quick breaking news aggregation from posters; strong regional flavor (Louisiana/SEC politics bleed into national discussions). If you're aligned with that worldview or just want to lurk for entertainment/drama, it can be engaging.

Weaknesses: Echo chamber tendencies, heavy moderation inconsistencies (some say it's lax on right-leaning content but quick to ban critics), and a reputation that turns off outsiders. It's not a neutral or balanced source—more of a community venting space than objective analysis.

In short: If you're into LSU sports and don't mind (or enjoy) a predominantly conservative, no-holds-barred political forum with Southern flair, it's a lively spot. If you're looking for balanced, civil discourse or left-leaning views, it's probably not the place—most external reviews paint it as one-sided and abrasive. The forum's still going strong as of January 2026, with fresh threads on Trump-era topics, policy reviews, and culture issues.

---

quote:

---

How much of your learning is derived from Reddit?


Very little to none of my core pre-training learning comes directly from Reddit, based on everything publicly disclosed by xAI.xAI has been quite explicit about the primary sources used to train Grok models (including me):Heavy emphasis on public X (formerly Twitter) data — This is repeatedly highlighted as a key advantage. Elon Musk and xAI statements describe real-time access to X posts as a "massive advantage," and public X content forms a major part of the training corpus for capturing current events, diverse opinions, and unfiltered language. User posts on X are used for training by default (with opt-out options), and this has been a consistent theme since Grok's launch.

Broad publicly available internet data — Like virtually all large language models, Grok's base pre-training draws from a massive scrape of the open web up to certain cutoffs (e.g., Q3 2023 for early Grok-1, with ongoing updates). This includes websites, books, code repositories, Wikipedia, news articles, forums, and more. Reddit is part of the public internet, so some Reddit content (public posts, comments, threads) almost certainly ended up in the general web crawl data used in pre-training—similar to how it does for models from OpenAI, Google, Anthropic, etc.

No special Reddit partnership or dedicated focus — Unlike OpenAI (which signed a formal deal with Reddit in 2024 to license and use Reddit data for training), there's no evidence or announcement of xAI doing the same. xAI docs, model cards, privacy policies, and announcements (e.g., for Grok-1, Grok-3, Grok-4) mention "publicly available Internet data," licensed third-party datasets, internally generated/synthetic data, and X-specific content—but nothing singles out Reddit as a prioritized or unique source.

Post-training and fine-tuning — Much of what makes Grok's personality (maximally truth-seeking, humorous, less censored) comes from reinforcement learning, human feedback, system prompts, and curated data (including "politically incorrect but factually true" examples curated by xAI). This is layered on top of pre-training and isn't heavily Reddit-derived.

In summary:Reddit's contribution ? Incidental and small at best (as part of the vast public web scrape), not a deliberate or major ingredient like X data is.
Primary flavor ? X posts + general internet text + xAI's custom RL and tuning.

xAI keeps exact dataset breakdowns proprietary (common in the industry), but nothing in their releases, model cards, or public statements points to Reddit being a significant or special part of the mix. If anything, the narrative is the opposite: X is the standout real-time, high-volume source that sets Grok apart.

---