Hi, I am a relatively new moderator and have a question regarding what user behavior can be considered spam.

According to this help article, the usage of generative AI tools and repeatedly posting unrelated/off-topic/link-farmed content that disrupts conversations within a community may qualify as spam.

As an example, we have a user named Rush20XX who very often initiates long comment threads and ALWAYS replies with AI-assisted paragraphs of text. Sometimes (~10-15% of the time) his answers can be helpful and useful (supported by upvotes), but more often they are just irritating, 3-to-6-paragraph blocks of verbose AI text with zero or negative upvotes.

At this point, I am starting to believe it might just be someone's Hermes AI agent responding most of the time. I'm not sure what the objective of these users is, but they can produce around 20 long comments a day, with most having a negative upvote score.

So, I am curious: is it acceptable to report such users for spam? Does this kind of behavior violate Reddit's Rule 2?

Desktop/Web.

• Lucky Pals Mill / Damage Check • Base Limits • The Taunt Mechanics • Blocking • Your Main Phase • The Quick Mechanic • Power vs Damage

What happens when I assign "Lamball – My First Pal" (ETD01-023) to "No Pals Beyond Sign" (EBP01-067) when the opponent has only 4-cost or greater Pals on their base? I have only 1 Taunt Pal (Lamball), so the opponent's attacking Pal should not be able to target anyone (even taunting Lamball thanks to his passive). Will such a situation prevent the opponent's 4+ cost Pals from attacking any of my Pals, forcing them to declare attacks only at structures or player HP? If not, why would the opponent's Pal be allowed to attack a non-Taunt Pal?

I believe this scenario is bugged at PalworldTCG.gg, and opponent could attack my other pals.

ChatGPT and Claude got their own laptops and credit cards and objective to wake Lemonade Stand in one week.

Link to comprehensive rules.

Link to Q&A about cards.

Hi, after publishing my prev post I found a lot more pack opening videos and fixed mistakes. Maintaining a post in Markdown syntax with manual recalculations and parsing card attributes became a pain. Therefore, I created a Google Sheets variant where I will keep track of new pack unwrappings (probably until reaching 100 packs or 1-2 SP+ cards).

Current stats/odds as of the post publish date and based on 97 packs are shown in the picture below (first sheet in Google Sheets).

Screenshot from Google Sheets with Distribution charts

Key points (chances are per card in a pack, not per pack):

• It looks like RR has an estimated ~5% chance.
• It looks like SR has an estimated ~2.5% chance.
• It looks like OSR has an estimated ~1% chance.
• It looks like SSP has an estimated ~0.25% chance.
• There was no SP present in the 48 packs, but I assume that SP has an estimated ~0.25-0.5% chance.
• You have an estimated ~10% chance to get a Lucky Pal per card in a booster pack.
• It looks like there is an almost guaranteed 1 card of (at least) OSR rarity per Booster Box.
• On average, every 2nd card from a booster pack will be a Pal, and every 10th card is a Lucky Pal.
• It looks like Colorless cards are the rarest (estimated ~4%) in booster packs, and most players would only hold Colorless cards from TD packs. The rest of the colors look balanced at an estimated ~24%.
• Every 7th card in a booster pack is guaranteed to be R+ (R/RR/SR/OSR/SP/SSP); AKA 1 crystal shiny/R+ card per booster pack is guaranteed.
• R and RR share the same holographic crystal shiny pattern.
• SR cards have a rainbow glassy foil + "out of bound/frame"/zoomed artwork.
• OSR cards have a diagonal shiny "textured rain" effect + special full-size artwork.
• SSP cards have "textured" outlines on the art, golden text, and special unique artwork.
• If you lucky you can pull 2 OSRs per box
• If you lucky you can pull 3 R+ cards in pack

Did DP-800 today (17.07.2026) thanks to free Fabric Data Days event voucher.

I just passed DP-800 and wanted to share what the exam was actually like, since it is a relatively new exam and there is not much info about it yet.

Exam structure

• The exam starts with a case study section (in my case, 9 questions), then moves into 41 regular questions (50 total).
• The last 3 questions are yes/no and not returnable or reviewable, meaning you cannot go back and change them once submitted. (I suspect these are always related to a T-SQL scripting solution for the stated objective.)
• You get 90 minutes, but questions have a lot of text. A typical question has 4 parts: context/setup, used features/services, problem statement, and the question itself, plus a (often large) schema for tables, an unfinished T-SQL script, a DAB schema, and so on. You may want to skip the setup paragraph and jump straight to the requirements, question, and answer options to save time.
• You have access to Microsoft Learn during the exam. I recommend using it heavily to verify your assumptions and doubtful answers, or to check docs for related topics and keywords. If needed, you can go to the Credentials page, find the DP-800 learning path, and search in modules during the exam.

Met Question Types

I have seen a mix of formats:

• Dropdown select
• Yes/No
• Radio button (single choice)
• Multi select
• Drag and drop (some items can be reused, others can be left blank)
• Case study based questions

No labs. No questions with screenshots or images.

Most questions (roughly 65%+) are T-SQL related, and from what I remember, mostly covered these topics:

• JSON handling
• GROUP BY and HAVING
• Window functions and PARTITION BY
• Indexes
• Stored procedures and functions
• Regex
• REST API calls from SQL
• Vector data types and search (embeddings, ANN)
• Data masking
• Encryption
• Permissions

I believe the exam had a good variety of T-SQL questions matching the topics listed in the official Study Guide.

Things That Felt Unexpected

• A few questions referenced stored procedures or features that were not mentioned in the official Microsoft Learn prep path. I had to search and read about them during the exam.
• One question asked whether float(32) is used under the hood for embedding or vector types. A slightly unexpected level of detail, but it was mentioned in one of the MS Learn modules.
• One question was about Azure AI Search indexes. This did not seem to be part of the official study plan or measured skills. (In my case, it was a question about a JSON config for an AI Search index.)
• Data API Builder (DAB) and CLI questions can be tricky. Be ready for things like which endpoint type DAB uses for Azure Static Web Apps, and CLI command usage for DAB and sqlpackage. Also know how to handle credentials safely.
• Expect 1-2 questions where you need to read an execution plan summary or warnings in text form, with no diagram or image.

GitHub Copilot Questions

These were the easiest part. All Copilot related questions could be answered with logic and general understanding, no deep memorization needed.

Tips

• Get comfortable using Microsoft Learn during the exam. You are allowed to check it, and you can even open specific modules from the learning path to validate your answer. Practice this before the exam so you are fast at finding things. Be ready to search for stored procedures, tables, or general SQL Server topics quickly.
• The Microsoft Learn learning path covers, in my opinion, about 90% of the exam questions. What it did not cover was mostly related to some specific system procedures and tables, plus one unexpected question about Azure AI Search index.
• Since some questions go beyond the official learning path, having hands-on T-SQL experience helps a lot, not just reading the docs. Doing the exercises from the learning path will help too.
• Time management matters. If you are running low, skip the setup text and focus on the actual question and requirements first.
• If you prefer videos, keep in mind that Youtube Channel Microsoft Reactor has few DP-800 prep videos and may release few more during next 2 weeks.

Happy to answer questions if anyone is prepping for this exam as far as I keep in mind exam specifics.

I just noticed that AZ-204 (Azure Developer Associate) will retire at the end of July 2026, and Microsoft is replacing it with AI-200 (Azure AI Cloud Developer Associate).

Table of retiring certifications and their replacements.

It looks like the AZ-204 page has even stopped rendering the MS Learning Path (I'm just getting an endless loading spinner/skeleton).

I was planning to take the AZ-204 later on, especially since it's frequently mentioned in other Reddit posts as holding a lot of weight for employment as an Azure Cloud Software Engineer, and is probably one of the most well-known certs (especially in combination with AZ-104).

I'm not sure if AI-200 will inherit the same credibility over the next few years as AZ-204, and its Learning Path looks a bit simpler. Even the current promotional courses for AZ-204 on Coursera (by Microsoft) seem to cover more cloud services, deeper material, and more practical experience...

I would love to hear your thoughts on these changes to the Azure certification retirement/replacement changes.