Rowan Cheung Profile picture
Apr 2, 2023 19 tweets 10 min read Read on X
ChatGPT can now browse the entire internet.

I spent my weekend analyzing the Twitter algorithm with help of the new GPT-4 browsing feature.

Here's EXACTLY how to optimize your Twitter posts for maximum virality:

🧵
1. Focus on Likes, then Retweets, then Replies.

Likes: 30x boost
Retweets: 20x boost
Reply: 1x boost
2. Include visual media.

Adding images or videos to tweets: 2x boost
3. Limit external links.

Avoid excessive external links, especially non-news or media sites.

Personally, I found links in the comments of your original tweet to not impact score (anecdotal).
4. Use trending topics.

Mention trending topics or news: 1.1x boost
5. Write complete sentences.

Use proper language, avoid making up words or misspellings, and form clear sentences to avoid penalties.
Before we move on, I write about how to use AI practically, just like how I used ChatGPT to summarize the Twitter algorithm.

Join 60,000+ others: therundown.ai/subscribe
6. Minimize hashtags.

Limit the use of multiple hashtags, as they can negatively impact performance.
7. Avoid offensive content.

Posting offensive or controversial content results in reduced reach.

-Mutes, blocks, unfollows, spam reports, and abuse reports negatively impact score.
-Getting mass unfollowed can get you shadowbanned for 3 months.
8. Engage with 'high-quality' accounts.

Interacting with low-quality accounts or accounts with poor follower-to-following ratios can lead to penalties.
9. Be mindful of audience and cluster.

Focus on content relevant to your audience and cluster to maximize engagement.

This means determining your niche, and only posting content within it.

In my case: it's AI.
10. Maintain a good Follower-to-Following ratio.

Following many more people than those who follow you hurts performance.

Use Twitter lists.
11. Pay for Twitter Blue.

Being verified with Twitter Blue (paid service) provides a boost to tweet reach.
12: Limit misinformation.

Avoid posting content categorized as misinformation, as it is heavily down-ranked.
13. Use English for UI and Tweets.

English user interface (UI) gives a 0.5x boost, and English tweets give a 0.2x boost.
Here's how I did it:

I fed ChatGPT/GPT-4 data from the open-sourced Twitter algorithm.

Then I fed it brilliant round-up threads I found on Twitter from these users:
@xerocooleth, @indexsy, @NFT_GOD, @aakashg0
@xerocooleth @indexsy @NFT_GOD @aakashg0 Then I simply asked ChatGPT to round up the most important points.

I extracted that information, and wrote this thread with added visual imagery (2x boost)
@xerocooleth @indexsy @NFT_GOD @aakashg0 If you can tell from this post and my profile @rowancheung, it's optimized for maximum virality.

If this post goes viral, then that means this experiment WORKED.

Surprisingly, I was already doing most of these things. My Twitter stats for March👇
@xerocooleth @indexsy @NFT_GOD @aakashg0 That's a wrap! One last thing- I spent all weekend curating this, and I need your help.

Please honor the gentlemen's agreement because I will STARVE without engagement dopamine:

1. Follow me @rowancheung for more
2. Like + Retweet the tweet below👇

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Rowan Cheung

Rowan Cheung Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @rowancheung

May 8
AI NEWS: Mistral just launched a new multimodal AI promising SOTA performance at much lower cost

Plus, more news from Meta, Anthropic, Microsoft, Google, Figma, and Netflix.

Here's everything you need to know:
Mistral just made two big reveals:

—Medium 3, a multimodal AI matching or surpassing 3.7 Sonnet, GPT-4o, and Llama 4 Maverick across benchmarks with 8x lower costs

—Le Chat Enterprise with corporate tools like Google Drive, agent building, and more
Meta AI dropped Meta Perception Language Model, an open & reproducible vision-language AI for challenging visual tasks

It can watch videos and extract details like what a person is doing in the content and how they are doing it
Read 10 tweets
May 7
AI NEWS: Google just released a new version of Gemini 2.5 Pro, taking coding and web development to the next level

Plus, more news from HeyGen, Lighttricks, Windsurf, FutureHouse, Ace Studio, and Recraft.

Here's everything you need to know:
Google released an early preview of Gemini 2.5 Pro I/O Edition, delivering:

—Enhanced performance for frontend and UI dev, code editing, and agentic workflows
—Top score on LM Arena and WebDev Arena, beating Claude 3.7 Sonnet
—Video understanding features
HeyGen dropped Avatar IV, an AI for expressive animations

With one photo and voice script, its audio to expression engine captures tone, rhythm, and emotion to generate facial motion

Also supports different subjects, camera shots, and formats
Read 10 tweets
Apr 28
AI NEWS: China's Baidu just dropped ultra-affordable "Turbo" models, taking on OpenAI and DeepSeek

Plus, more news from OpenAI, Liquid AI, xAI, and Moonshot AI!

Here's everything you need to know:
Baidu dropped faster and cheaper ERNIE 4.5 Turbo and X1 Turbo models

4.5 Turbo, with upgraded multimodality, costs 11c and 44c per million I/O tokens (0.2% of GPT-4.5).

X1 Turbo comes at 14c/55c per million I/O tokens, topping DeepSeek R1 and V3
OpenAI released a new lightweight version of Deep Research to meet user demand

—Powered by o4-mini and ‘nearly as intelligent’
—Expanded usage limits but shorter responses
—Available on all tiers, including free users
Image
Read 9 tweets
Apr 22
TODAY'S AI NEWS: Google DeepMind CEO Demis Hassabis just said AI can end all disease!

Plus, more news from Anthropic, Windsurf, ElevenLabs, Play AI, and TextArena.

Here's everything you need to know:
Demis Hassabis appeared on 60 Minutes, sharing new insights into AI:

—AGI in 5-10 years, with conscious AI likely in the future
—AI fast-tracking drug discovery, potentially eradicating all disease within the next decade
—AI ensuring radical abundance
Anthropic just mapped Claude's morality!

The co analyzed 300K+ chats and found that the AI exhibits 5 key values, including practical and protective

This will improve our understanding of AI as it becomes more and more crucial to real-world decisions
Image
Read 9 tweets
Apr 17
TODAY'S AI NEWS: OpenAI just released o3 and o4-mini, its smartest models yet

Plus, more news from Google, Anthropic, Cohere, Microsoft, Kling AI, and more.

Here's everything you need to know:
OpenAI dropped the new o3 and o4-mini reasoner models

o3 pushes SOTA performance across coding, math, science, and multimodality, while o4-mini offers fast, cost-efficient performance

Both have agentic access to ChatGPT tools and can "think with images"
Google is expanding Gemini Live's Project Astra capabilities—to all Android users!

This will enable users to engage with real-time visual AI and have multilingual conversations about anything seen and heard via their phone's camera or screen share
Read 11 tweets
Apr 15
TODAY'S AI NEWS: OpenAI just dropped three GPT-4.1 models for devs—API-only and packing a jaw-dropping 1M-token context window.

Plus, more news from Google, Hugging Face, ByteDance, SSI, and LM Arena.

Here's everything you need to know:
OpenAI released GPT-4.1, 4.1 Mini, and the ultra-fast 4.1 Nano—all designed for devs

— Each model beats GPT-4o and 4o mini on dev tasks
— 1M token context windows
— GPT-4.1 scored 55% on SWE-Bench Verified
— Starting at $0.10/0.40 per million I/O tokens
Google released DolphinGemma, an AI that can generate dolphin vocalizations

The system analyzes dolphin whistles and sounds to identify patterns and predict subsequent sounds!

Set to be open-sourced soon!
Read 9 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(