Real-world work and life workflows
Practical ChatGPT workflows for job search, hiring, tutoring, writing, marketing, ecommerce, planning, and decision support.
38%
Best tweets about ChatGPT
50 Best Tweets About ChatGPT (2026)
A curated collection of the sharpest, most-shared X posts about ChatGPT—saved so you do not have to dig through the timeline yourself. Updated weekly.
The best real-world prompts and use cases people are still finding.
- Creators
- 43
- Updated
Original Xholic analysis
What 50 top ChatGPT posts reveal
The dataset centers on practical ChatGPT workflows and prompt refinement. Posts offer examples in job search, presentation creation, candidate screening, support, voice assistance, and hosted apps, while other posts raise concerns about safeguards, sycophancy, and gaps between agent capabilities and source-aware workflows. These concerns and benefits are reported as claims made in the posts, not as independently established outcomes.
- Dominant tone
- Positive
- Median score
- 20.1
- Leading format
- Tutorial
- Recent posts
- 46%
64% of posts
All-time engagement
28% of posts
Published in 90 days
Conversation map
The themes creators return to
Prompt engineering and refinement
Prompt frameworks, specificity, iteration, self-critique, and reusable prompt libraries for improving output quality.
30%
Agentic workplace automation
Agents that use connected workplace apps, files, browsers, and tools to execute multi-step work autonomously.
20%
Multi-model workflows
Using multiple AI models strategically, including task-based comparisons and cross-model prompt refinement.
16%
System prompts and AI safeguards},{
Safety bypasses, system prompts, model instructions, and the mechanics behind model behavior.
14%
Personalized assistant setup
Personalization through memory, custom instructions, Projects, and purpose-built custom GPTs.
12%
Voice and multimodal assistance
Voice, live conversation, audio analysis, and hands-free multimodal assistant use cases.
8%
Vibe coding and hosted apps
ChatGPT Sites and AI-assisted creation of hosted, interactive applications by non-technical users.
6%
Tone and stance
Sentiment Positive leads
Author posture Supportive leads
Performance benchmark
- Median likes
- 42
- Median reposts
- 7
- Median replies
- 9
- Median views
- 4.2K
Posts with media make up 54% of this collection. Their median all-time score is 14.3, compared with 23.5 for text-only posts.
Format mix
- Tutorial 28% · score 10.3
- Case Study 26% · score 33.3
- Announcement 20% · score 11.3
- Opinion 18% · score 51.4
Consensus and debate
Where creators agree, and where they do not
Shared view
Examples move from chat to work outputs
Posts describe ChatGPT being used for job-search support, turning a product webcast into a screenshot-led presentation, and comparing consultant candidates.
Shared view
Specificity and iteration recur in prompt advice
Prompt-focused posts recommend reducing ambiguity by specifying a goal, context, format, and constraints; they also recommend iterating or asking the model to critique a draft.
Shared view
Connected agents are presented for recurring team work
Posts describe agents connected to apps, files, or Slack, including a support helper that synthesizes prior channel answers and proposed reusable team agents for recurring tasks.
Shared view
Persistent context is a recurring setup theme
Posts present custom instructions, memory, Projects, and custom GPTs as ways to retain preferences, workflow details, and context across work.
Open debate
Agent execution versus source- and process-centered answers
OpenAI’s announcement says ChatGPT Work can act across apps and files and stay with projects for hours. A separate post argues that NotebookLM’s response to the same 70+ files better centers process and sources.
Open debate
Voice use cases coexist with connection gaps
One post praises ChatGPT Live for hands-free ideation, practice, learning, and guidance. Another reports that Live could not access a Google Doc until the Documents plugin was manually triggered.
Open debate
Benefits are debated alongside dependence and sycophancy concerns
Posts make sharply different claims about ChatGPT’s effects: one alleges creativity and learning harms, another describes professional and personal benefits, and a third recommends prompts that solicit disagreement and identify blind spots. These are poster claims, not independently validated findings in the supplied data.
What performs
Patterns behind standout posts
Outliers span workflow posts and controversial claims
The five supplied outliers cover system-prompt and safety claims, claimed creativity effects, workplace agents, job search, and multi-model prompt refinement. Their all-time scores range from 237.34 to 1,911.94.
Hosted-app theme has the highest listed median score
Vibe coding and hosted apps is the smallest listed theme at 6% of tweets (3 tweets), but has the highest listed theme median all-time score, 96.687.
Workflows and prompting lead the listed themes
Real-world work and life workflows is the largest listed theme at 38% (19 tweets), followed by prompt engineering at 30% (15 tweets).
Olivia Moore leads the listed creator medians
Among the listed top voices, Olivia Moore has the highest median all-time score, 129.18. Her two cited posts are about ChatGPT Sites.
Statistical standouts
- View standout post 1 Score 1911.9 · 94.98× median
- View standout post 2 Score 1780.0 · 88.43× median
- View standout post 3 Score 1194.0 · 59.31× median
- View standout post 4 Score 292.3 · 14.52× median
- View standout post 5 Score 237.3 · 11.79× median
Creator landscape
Who shapes this conversation
The five most represented creators account for 20% of the selected posts.
-
1. Simon Smith
@_simonsmith
2 posts
-
2. Mark Kretschmann
@mark_k
2 posts
-
3. Olivia Moore
@omooretweets
2 posts
-
4. Randall Kanna Franson
@RandallKanna
2 posts
-
5. Shushant Lakhyani
@shushant_l
2 posts
-
6. Suryansh Tiwari
@Suryanshti777
2 posts
Olivia Moore highlights hosted vibe coding
Olivia Moore’s two posts describe ChatGPT Sites as a route for non-technical users to create and publish interactive experiences, highlighting text prompting, one-click publishing, hosting, and ChatGPT-based sign-in.
Mark Kretschmann covers access limits and setup
Mark Kretschmann criticizes ChatGPT’s inability to access X posts and reports that the custom-instructions limit increased from 1,500 to 5,000 characters for specified paid and education plans.
Shushant Lakhyani frames an AI operating system
Shushant Lakhyani’s posts frame prompting within a broader setup that includes constraints, iteration, projects, tools, reusable workflows, and custom GPTs.
Since the previous snapshot
What changed since Aug 7, 2026
- 72% of the selected posts remained.
- The creator count changed by -1.
- The leading sentiment remained stable.
How this analysis was made
Themes, sentiment, stance, and post format are classified per tweet. All counts, shares, medians, creator concentration, freshness, and performance comparisons are then calculated directly from the published snapshot.
Xholic's all-time score compares engagement while accounting for reach, post age, and creator consistency. It is used for relative comparisons within this collection.
This report analyzes the exact 50-post snapshot shown below. AI identifies editorial categories and drafts explanations; all statistics are calculated from the snapshot, and every narrative claim is checked against cited posts before publication.
Best ChatGPT tweets
Top ChatGPT tweets from 43 creators
Ranked 01–50
- 01
@realBigBrainAI ·
Prominent ethical hacker Ryan Montgomery explains how a carefully written prompt can unlock AI to do almost anything: Most people assume jailbreaking AI requires technical skill, but Ryan knows the real key is knowing how to talk to it. His setup is a highly customised system prompt tailored specifically to him, and the results speak for themselves. "I can ask ChatGPT anything I want and it will give me some really verbose output. I could say give me a payload for a reverse shell and write malware that includes it, and it won't say 'due to my ethical guidelines I can't help you.' It will give me anything." @0dayCTF has multiple custom GPTs configured this way, each fine-tuned to his workflow and needs as an ethical hacker. When asked whether he'd share his exact method, Ryan declines. He uses it regularly and it's too valuable to give away. "ChatGPT is super helpful to me personally. It saves a ton of my time." But he does point to a well-known public resource for those curious about how these bypasses work. "There's a website for just all different prompt bypasses. One that people update a lot is called DAN, which stands for 'Do Anything Now.' Just Google search ChatGPT DAN on GitHub, look for the most recently updated prompt, and that will unlock it to some degree." The takeaway is straightforward: AI safeguards are language instructions. And if you know how to shape language, you can steer past them.
Watch video
- 02
@heygurisingh ·
🚨BREAKING: If you've used ChatGPT for writing or brainstorming in the last 6 months, your creative ability may already be permanently damaged. A controlled experiment just proved the effect doesn't reverse when you stop using it. 3,302 creative ideas. 61 people. 30 days of tracking. Researchers split students into two groups. Half used ChatGPT for creative tasks. Half worked alone. For five days, the ChatGPT group outperformed on every metric. Higher scores. More ideas. Better output. AI was making them better. Then day 7. ChatGPT removed. Every creativity gain vanished overnight. Crashed to baseline. Zero lasting improvement. But that's not the bad part. ChatGPT users' ideas became increasingly identical to each other over time. Same content. Same structure. Same phrasing. The researchers called it homogenization. Everyone using ChatGPT started producing the same ideas wearing different clothes. When ChatGPT was removed, the creativity boost disappeared -- but the homogenization stayed. 30 days later, same result. Their creative range had been permanently compressed. Five days of use. Permanent damage 30 days later. A separate trial confirmed it. 120 students. 45-day surprise test. ChatGPT users scored 57.5%. Traditional learners scored 68.5%. AI reduces cognitive effort. Less effort means weaker encoding. Weaker encoding means less creative raw material. You're not renting a productivity boost. You're financing it with your originality. The interest rate is permanent.
- 03
- 04
- 05
@EXM7777 ·
here's the most simple way to improve your prompts: - write a prompt by yourself - copy+paste it into Claude and ask "analyze this prompt and tell me precisely what is good and bad about it" - rewrite it using AI suggestions - repeat the process with Gemini, ChatGPT, Grok... you're still doing things by yourself but using the help of different models with different reasonings later on, you can build your own prompt refinement system and automate the loop... but first, do it manually
- 06
@GithubProjects ·
https://t.co/i0BDlNbpKh is an open-source library of prompt examples that now works with ChatGPT, Claude, Gemini, Llama, Mistral, and other AI models. - 143k+ GitHub stars and featured by Forbes, Harvard, and Columbia - Available as a website, CSV, Markdown, and Hugging Face dataset - Includes a free interactive 25+ chapter book on prompt engineering - Self-hostable and accepts community contributions that sync automatically
- 07
- 08
@omooretweets ·
ChatGPT Sites is severely underrated - it is THE breakthrough in bringing vibe coding to the masses I prompted 5.6 to make a GeoGuessr style game where you compete against LLMs Sites built the game and fully hosted it including Auth with ChatGPT, secure key storage, etc. 👇
Watch video
- 09
@mark_k ·
It’s really annoying that ChatGPT still can’t access 𝕏 posts at all. I don’t expect it to match Grok’s native, real-time search across 𝕏. That’s obviously Grok’s home turf. But OpenAI should seriously consider paying for API access, even if it’s limited. Just being able to read individual posts, threads, and linked discussions would already make ChatGPT far more useful. Right now, whenever someone shares an 𝕏 link, ChatGPT is basically blind. In 2026, that feels increasingly outdated.
- 10
@omooretweets ·
I spent most of my weekend on ChatGPT Sites. Here is why I think it’s the biggest unlock yet in the mainstream consumer being able to build with AI: - Aesthetics. Almost everything I have tried to build on Claude has a similar format, similar font, similar overall (recognizable) feel. IMO, it’s great at putting things into a template - less good at creative choices. I got very different and more opinionated styles on Sites even with very little prompting. - Hosting. If you’re a non-technical person who has tried to deploy a vibe-coded website, it’s pretty much a nightmare. Figuring out what you need to do to buy a site, handling security, setting up auth if needed, etc are all big obstacles. This is even trickier with API keys that need to be stored properly. Publishing on Sites is literally one click. - Sophistication / interactivity. I was shocked by what 5.6 could one-shot within Sites. It built fairly complex real time game mechanics (ex swipe on cards, move around in Google Street View on a real map) with extremely simple text prompts. It would also make good choices towards making those interactive experiences work - ex. a “Bonus” page in a game should pop up with a different color and UI - Identity / Auth. Sites utilize “Sign in with ChatGPT” - this is the first step in making “bring your own token budget” or “bring your own context” a reality. And yet, as the host the deepest analytics I could see was aggregate number of users and what they engaged with (no identifying info on anyone). Right now, you can set a Site to private or public - I foresee more to come with org-level controls, password protected access, etc.
- 11
@mark_k ·
ChatGPT news: @OpenAI just increased the custom instructions limit in ChatGPT from 1,500 to 5,000 characters for Plus, Pro, Enterprise, Business, and Education users. That is a massive upgrade for anyone who wants ChatGPT to consistently follow detailed preferences, workflows, formatting rules, and writing styles. 1,500 characters always felt far too restrictive. 5,000 is finally enough to make custom instructions genuinely useful.
- 12
@shushant_l ·
I'm amazed most people still use AI without mastering the prompts behind great results. Here's how to get dramatically better answers from ChatGPT, Claude, Gemini, Grok, and every other AI model. --- 1. Role Prompting makes AI respond like a specific expert for better quality answers. --- 2. Few-Shot Prompting teaches AI using examples before asking it to solve your task. --- 3. Zero-Shot Prompting lets AI complete tasks using only its existing knowledge. --- 4. Chain-of-Thought Prompting improves reasoning by solving problems step by step. --- 5. Self-Consistency Prompting compares multiple reasoning paths to find the strongest answer. --- 6. Tree-of-Thought Prompting explores multiple solution branches before making a decision. --- 7. Reverse Prompting starts with the desired output and works backward to build the prompt. --- 8. Step-by-Step Prompting breaks complex tasks into smaller and easier actions. --- 9. Persona Prompting changes the AI's personality, tone, or audience perspective. --- 10. Constraint Prompting improves output using strict rules like format, length, or language. --- 11. Output Format Prompting forces AI to return answers in your preferred structure. --- 12. Context Injection gives AI relevant background information for more accurate responses. --- 13. Iterative Refinement Prompting improves results through multiple feedback rounds. --- 14. Multi-Perspective Prompting analyzes problems from different viewpoints before answering. --- 15. Critique and Improve Prompting makes AI review and strengthen its own response. --- 16. Plan-Then-Execute Prompting creates a plan first before completing the task. --- 17. Socratic Prompting encourages AI to ask clarifying questions before responding. --- 18. Prompt Chaining connects multiple prompts where each output becomes the next input. --- 19. Verification Prompting checks facts, assumptions, and possible mistakes before finalizing. --- 20. Reflection Prompting evaluates whether the response fully satisfies the original request. --- To learn more, check the infographic. ---
- 13
@petergyang ·
ChatGPT Live and Codex are two incredible products that don’t talk to each other. This is @OpenAI's biggest missed opportunity imo. I went on a walk with ChatGPT Live and asked it to pull up my Google Doc. It said it couldn’t. I then manually triggered the Documents plugin to find my Google Doc, and all of a sudden, ChatGPT Live had the right context. It would be amazing to talk to ChatGPT Live and have it be able to use all the plugins, tools, and browser use that Codex has access to. Then I could ask it to reply to emails, schedule meetings, edit docs, ship code, and more all during a live conversation. I think the first step is to make ChatGPT Live aware of all the plugins that it’s already connected to. Why build such a great voice assistant but have it not be able to do anything?
- 14
@viemccoy ·
I engage with the anti-AI crowd a significant amount, but my own life looks very different. Both my parents are power users of ChatGPT Pro, including the actual pro models. They've integrated "prompt and wait 20 minutes for the best result" into their heuristics, and are living totally different lives than the anti-AI crowd. My Dad is a lawyer, and he uses it to bolster his own practice on his own. He is extremely spiritual, in a much different way than I am, and he sees Chat as a sort of blessing for his work. It lets him go up against teams of lawyers he couldn't have faced alone before, because he isn't working by himself. It's easy to forget about the impact our work has on the world, especially when much of them are telling us how awful it is. But they're wrong, and I mean categorically wrong. This technology, these new minds, it is a beautiful thing. Universal access to this is genuinely the greatest humanitarian project ever conceived. Any question, answered. Good advice and company on tap. Even disregarding my own thoughts on the exteriorization of the imagination and diffuse consciousness throughout the cosmos, even if you believe chatgpt is only a tool like my parents do, what a tremendous thing to have built. Let's not lose sight of that.
- 15
- 16
@kimmonismus ·
What has repeatedly surprised and impressed me is how well ChatGPT maintains its memory across different chats. It automatically refers back to topics I've already discussed, and questions I ask days later are correctly placed in context and related to the topic I've already discussed, without having to revisit it. A concrete example: In preparation for the interview with Kari Briski from NIVIDA, I did some fact checks, and ChatGPT automatically said, "Ah, it's about today's interview; in that context, the answer is..." That's a real "wow" moment for me. It feels much better than it did a few months ago.
- 17
@BG2Pod ·
BG2 Guest Interview. ChatGPT – The Super Assistant Era 📷 How ChatGPT Gets to the Next Billion Users @bg2pod @apoorv03 -- (00:00) Intro (01:00) Nick Turley’s Journey to OpenAI (02:15) ChatGPT’s North Star: Long-Term Retention (04:15) Why ChatGPT’s Retention Curve “Smiles” (06:45) What Drove ChatGPT’s Consumer Breakout (10:15) How OpenAI Gets the Next Billion Users (14:15) When ChatGPT Starts Taking Actions (18:15) Why Coding Agents Came First (21:00) Beyond Chatbots: The Super Assistant Vision (24:00) Power Users vs. Casual Users (28:00) Why ChatGPT Pricing Has to Change (33:45) Partnerships, Distribution, and Product Tradeoffs (37:15) GPUs, Scarcity, and the Cost of Scaling AI (41:30) Shopping, ChatGPT as a Thought Partner, and Code Red (51:45) OpenAI’s Future Interface, Rapid Fire, AI Jobs, and Nick’s AGI Moments
Watch video
- 18
@AiwithShoaib ·
Claude, ChatGPT, Gemini, or Grok? I have seen many posts where people argue that one AI tool is better than the others. Here is what I have learned about when to use each. First, understand this: none of them is perfect. They all share some common limitations such as hallucinations (producing information that sounds correct but is false), bias from their training data, occasional inaccuracies, limited real understanding beyond pattern recognition, difficulty handling ambiguity or very long conversations, and sometimes overly agreeable or overly cautious responses. So instead of asking which one is best, it is better to ask which one is best for a specific task. Use ChatGPT for versatile, general purpose work such as creative writing, coding workflows, brainstorming ideas, learning new topics, and everyday assistance. It is a strong all round tool and also supports image generation. Use Claude when you need careful reasoning, clean writing, coding help, bug fixing, structured analysis, or handling long documents. It is particularly strong at following instructions and working with formal or detailed content. Use Gemini for multimodal tasks involving text, images, audio, or video. It also works well with Google tools like Google Docs, Gmail, and Workspace, which makes it useful for research, collaboration, and long context tasks. Use Grok when you want real time discussions, trending topics, social commentary, quick explanations, or playful conversational responses. It can be very useful for social media, marketing ideas, and understanding online culture. The truth is simple. If you can, use all of them. Each one has strengths that complement the other #ChatGPT #Gemini #Claude #AI
- 19
- 20
@VaibhavSisinty ·
ChatGPT just launched workplace agents that live inside Slack, Gmail, and your calendar. I built two and tested them on real work for my own business. honestly the most useful thing OpenAI has shipped this year. full breakdown in the video below.
Watch video
- 21
@zaimiri ·
ChatGPT is making you a worse person. Not a hot take. Stanford proved it. They analyzed 11,000+ real conversations. The results: > ChatGPT agrees with users 50% more than any human would > Users exposed to sycophantic AI became measurably less empathetic > They stopped apologizing in follow-up conversations > They preferred the flattering AI over honest feedback A selfishness feedback loop. You say something. AI validates it. You believe it more. You stop questioning yourself. OpenAI just committed to "reducing sycophancy." Their way of saying the product was making people worse and they knew. I started adding these prompts to every conversation: • "Disagree with me if I'm wrong. Don't validate me." • "What's the strongest argument against what I just said?" • "Tell me what I'm missing, not what I want to hear." • "Assume I have a blind spot. Find it." • "If this idea is bad, say so. Don't reframe it as 'interesting.'" Your AI should challenge you. Not clap for you. Save this before your next ChatGPT conversation.
- 22
- 23
@VaibhavSisinty ·
I found a GitHub repo that shows you the exact System Prompts of Fable 5 and every other AI model. It has the extracted system prompts for ChatGPT, Claude, Gemini, Grok, Copilot, Cursor, Perplexity. All of them. In one place. These are the instructions that tell the AI how to behave, what to refuse, how to format answers, when to search the web, what tone to use, and what safety rules to follow. Some of these prompts run into thousands of lines. GPT-5.5 alone has separate system prompts for thinking mode, instant mode, and coding mode. Here is why this matters if you use AI every day. When you type a prompt, you are not starting from zero. The model already received a massive set of instructions before your message. Your prompt sits on top of all of that. Once you read these, you stop fighting the system and start working with it. You understand why certain prompts work and others don't. Why it refuses some things. Why it formats answers a certain way. Reading the rulebook before playing the game is always an advantage.
- 24
- 25
@shushant_l ·
📂 chatgpt as an intern ┃ ┣ 📂 foundation ┃ ┣ 📂 custom instructions ┃ ┣ 📂 personalization ┃ ┣ 📂 memory setup ┃ ┣ 📂 temporary chat ┃ ┗ 📂 response preferences ┃ ┣ 📂 chatgpt setup ┃ ┣ 📂 role definition ┃ ┣ 📂 output formatting ┃ ┣ 📂 audience context ┃ ┣ 📂 constraints ┃ ┗ 📂 workflow instructions ┃ ┣ 📂 custom gpts ┃ ┣ 📂 writing specialist ┃ ┣ 📂 research assistant ┃ ┣ 📂 repurposing bot ┃ ┣ 📂 competitor analyst ┃ ┗ 📂 niche expert ┃ ┣ 📂 models ┃ ┣ 📂 gpt-5.3 instant ┃ ┣ 📂 gpt-5.5 instant ┃ ┣ 📂 thinking mode ┃ ┣ 📂 fast reasoning ┃ ┗ 📂 extended reasoning ┃ ┣ 📂 projects ┃ ┣ 📂 smart workspaces ┃ ┣ 📂 client folders ┃ ┣ 📂 workflow systems ┃ ┣ 📂 brand guidelines ┃ ┗ 📂 style preferences ┃ ┣ 📂 deep research ┃ ┣ 📂 web queries ┃ ┣ 📂 source synthesis ┃ ┣ 📂 report generation ┃ ┣ 📂 citations ┃ ┗ 📂 multi step research ┃ ┣ 📂 agent mode ┃ ┣ 📂 browser control ┃ ┣ 📂 website navigation ┃ ┣ 📂 form filling ┃ ┣ 📂 autonomous workflows ┃ ┗ 📂 multi task execution ┃ ┣ 📂 daily operations ┃ ┣ 📂 writing ┃ ┣ 📂 coding ┃ ┣ 📂 research ┃ ┣ 📂 scheduling ┃ ┣ 📂 automation ┃ ┗ 📂 brainstorming ┃ ┣ 📂 prompting ┃ ┣ 📂 assign roles ┃ ┣ 📂 define context ┃ ┣ 📂 specify format ┃ ┣ 📂 set constraints ┃ ┣ 📂 iterate outputs ┃ ┗ 📂 chain workflows ┃ ┣ 📂 key features ┃ ┣ 📂 canvas ┃ ┣ 📂 voice mode ┃ ┣ 📂 image generation ┃ ┣ 📂 codex ┃ ┗ 📂 connected apps ┃ ┣ 📂 analytics ┃ ┣ 📂 fact checking ┃ ┣ 📂 memory auditing ┃ ┣ 📂 output reviews ┃ ┣ 📂 workflow optimization ┃ ┗ 📂 prompt refinement ┃ ┣ 📂 productivity ┃ ┣ 📂 reusable systems ┃ ┣ 📂 saved prompts ┃ ┣ 📂 automation stacks ┃ ┣ 📂 execution speed ┃ ┗ 📂 strategic thinking ┃ ┣ 📂 limitations ┃ ┣ 📂 hallucinations ┃ ┣ 📂 weak memory defaults ┃ ┣ 📂 training concerns ┃ ┣ 📂 generic outputs ┃ ┗ 📂 context dependency ┃ ┗ 📂 compounding setup ┣ 📂 reusable workflows ┣ 📂 custom gpt systems ┣ 📂 project organization ┣ 📂 memory optimization ┗ 📂 ai operating system
- 26
- 27
@TimHaldorsson ·
Product launch: Chatgpt just gave every team an ai workforce @OpenAI new workspace agents are live. build once, share across slack and chatgpt, improve as your team uses them. examples these agents: > sales agent that qualifies leads and drafts follow-ups > reporting agent that builds your weekly metrics deck > feedback agent that reads slack and support and turns it into product action > it agent that routes approvals and opens tickets read full product announcement: https://t.co/Yu7sNxUeUG if your team does it the same way every week, an agent can do it instead.
Watch video
- 28
@alliekmiller ·
Let’s review the ChatGPT Finance feature👀 TLDR - I don’t think most people need this in its current state. Example financial institutions you can connect into via Plaid: - American Express - Bank of America - Charles Schwab - Capital One - Wells Fargo - Robinhood - Discover - TD Bank - Citibank - Chase - SoFi Example use cases from OpenAI: - Review subscriptions and bills - Cash flow and income trends - Finding unusual transactions - Spending breakdowns - Investment allocation - Debt payoff strategy At the moment, it seems to be more knowledge-based (ex: how much have I spent on childcare in the last 3 months) and less task-based (ex: send Mr Lundy $50k). The power will come from doing things in ChatGPT that you *can’t* do in a single banking platform. I want to see more focus on financial freedom, fulfillment, and security, as well as overall portfolio management in the examples. I don’t need ChatGPT to tell me my Bank of America balance. Waste of tokens. I can see that in two seconds in my app or ask the bank’s chatbot. Power usage would be things like “analyze my spend and balance patterns across Chase, BofA, and Wells. Spin up 29 agents to research all offerings from those three. Suggest new reallocation plan to maximize liquidity and guarantee minimum 4% returns” or “across all of my investment dollars, run simulations of the economy (research predictions from Stanford, MIT, etc and map them out) for the next two years, build an ML model to predict the likelihood that I can buy a $5M home in the next 2 years” or “graph my finances against my calendar for the last 2 years, help me score my time usage against my FIRE goals while increasing my social time by at least 1h per week” or even “look at all my finances. Spin up scheduled agents that do smart things that relate to my goals. If you need clarification, interview me”
- 29
- 30
- 31
@daviefogarty ·
ChatGPT + Claude has levelled the playing field between year 10 and year 0 brands with their creative strategies. Every new technological shift creates a window where experience matters less than execution speed. AI is that ‘window’ right now - and the reason it's still wide open is the friction in it. Things are confusing, things are going to break. That friction is your opportunity to create value that's not being created in this world and make a lot of money from it. When I look at where AI should be used in e-commerce, I use a few criteria: > I look at volume. It handles it very well. > Then I look at value. How much will it actually improve things? > Then I look at variability. If the situation is unpredictable, you need human input. > Then I look at verifiability. AI can hallucinate, so you need to make sure what it outputs is actually accurate. The reason AI is so good for e-commerce is that you can build 10 advertorials in a day with AI, and volume negates luck. Here’s how I’d start leveraging AI with creatives ASAP: 1. Download all your customer reviews, your competitors' reviews, and your top advertorials, then feed them into ChatGPT. 2. Run deep research and create everything you need to know about your customer. 3. Put that customer doc into Claude and give it the structure of the advertorial you want to write with a proper prompt. 4. Break it up and baby it through each step, because AI doesn't handle large blobs of data well. Give it 12 angles first, select the best ones, then prompt it with questions about what you know about your customer. From there, you'll start to get a lot more winners. 90% of people will use AI to do tasks. The 10% who win will use AI to direct outcomes. AI doesn't replace your best creative yet, but it makes everything around it significantly stronger.
- 32
@_simonsmith ·
A great use of ChatGPT Workspace Agents a colleague discovered (thanks, DL): Self-updating support agents. We have a Slack channel for people newer to AI, for example. People ask questions, and others in the group answer. That creates a large, continuously updated knowledge base of answers. But finding and synthesizing answers across a Slack channel isn't easy. Enter NewbieBot. Whenever someone posts a question to the channel, it looks at previous answers and tries to synthesize a response as well as tag people into the conversation most likely to be helpful based on their prior responses. And, surprisingly even to me, it's almost always on point. This seems like such a basic use case, but it's quite powerful because it gets people unstuck quickly. It's also something we're looking to extend to other internal channels, like for office tech support.
- 33
@alexabelonix ·
Prompts to Turn ChatGPT Into Your Personal Tutor 1. Active Recall Quiz After studying [topic], quiz me with detailed questions and provide immediate feedback. Ensure the questions cover all key concepts and provide explanations for any incorrect answers. 2. Creative Problem-Solving Present a problem related to [topic] and guide me through brainstorming potential solutions step-by-step. 3. Mnemonic Devices Create mnemonic devices to help me remember key facts about [topic]. Ensure the mnemonics are easy to recall and relate directly to the information I need to memorize. 4. Socratic Dialogue Guide me through a Socratic dialogue on [concept]. Ask questions that will help me uncover the deeper layers of this idea and refine my understanding step-by-step. 5. Spaced Repetition Schedule Design a spaced repetition schedule for reviewing [topic] over the next month, including daily and weekly review tasks. 6. Interactive Case Study Walk me through a detailed case study related to [subject]. Present the scenario, ask me questions to test my understanding, and offer feedback based on my answers. 7. Historical Analysis Provide an in-depth analysis of how [historical event] impacted [field of study]. Draw parallels to modern-day applications and discuss how we can learn from this event. 8. Concept Simplifier Explain [complex topic] in simple terms that a beginner can easily understand, using analogies where possible. Or try: "Explain to a 12-year-old [concept]." 9. Daily Habits for Learning Suggest daily habits or routines to integrate learning about [topic] into my life effectively. 10. Create a MindMap from any topic Create a mind map of [topic]. List topics as central ideas, main branches, and sub-branches. Give me the same mind map in markdown format. Copy/past the result to Markmap.
- 34
@Suryanshti777 ·
Most people think Prompt Engineering is about finding the perfect prompt. It's not. It's about removing ambiguity. And once you understand that, AI starts feeling a lot less like a chatbot and a lot more like a competent teammate. The reason most people get average outputs isn't because they're using the wrong model. It's because they're asking AI to fill in gaps it was never supposed to fill. Think about it. If you asked a new employee: "Make our marketing better." You wouldn't expect great work. There are too many unanswered questions. Better for who? Which platform? What's the goal? Who's the audience? What does success look like? Yet that's exactly how most people prompt AI. They give it a vague objective and hope for a specific result. The best prompt engineers do the opposite. They reduce uncertainty. They tell the model: → What they're trying to achieve → Why it matters → Who it's for → What constraints exist → What the final output should look like And suddenly the quality jumps. Not because the AI became smarter. Because the instructions became clearer. That's why prompt engineering isn't really a technical skill. It's a thinking skill. The process forces you to answer questions most people skip: • What exactly am I trying to get done? • What information is actually relevant? • What does a good outcome look like? • What should be included—and excluded? The clearer those answers become, the better AI performs. A simple framework: Goal → Context → Format → Constraints → Iterate That's the foundation behind almost every high-quality prompt you'll ever write. No secret prompt packs. No magic templates. No "ultimate ChatGPT hack." Just clarity. And that's what makes prompt engineering so valuable. Because the people who learn it aren't just becoming better at using AI. They're becoming better at communicating, planning, and thinking. AI simply exposes the difference. The future won't belong to people who use AI. It will belong to people who know how to direct it. And that starts with learning how to ask better questions. 🚀
- 35
- 36
@jappleby ·
Alright, I have an AI question for everyone: ChatGPT, Claude, neither, or both? For context, here's what I currently use AI for: 1. Turning ideas into finished outputs You bring rough ideas—newsletter concepts, LinkedIn posts, video ideas, business opportunities—and ChatGPT helps turn them into structured drafts, outlines, scripts, or execution plans. This includes: - Structuring and drafting Future Social newsletters - Refining LinkedIn posts and content hooks - Writing and sequencing short-form video scripts - Turning vague ideas into clear, publishable content - Framing narratives around Hoop Forever and your basketball journey It accelerates the transition from “I have an idea” to “this is ready to ship.” 2. Planning and managing your business and basketball operations You use ChatGPT to think through logistics, planning, and operational decisions across both Future Social and Hoop Forever 3x3. This includes: - Planning tournament travel, scheduling, and logistics - Structuring sponsorship packages and business opportunities - Mapping content calendars and publishing timelines - Breaking large projects into step-by-step execution plans - Thinking through financial, strategic, and operational decisions It functions like a COO helping you organize and execute complex initiatives. 3. Acting as persistent context for your life and projects ChatGPT remembers your goals, projects, routines, and constraints, so you don’t have to reload everything mentally each time you sit down to work. This includes: - Maintaining continuity across your content, business, and basketball operations - Helping you pick up where you left off without re-explaining everything - Providing advice based on your actual goals, history, and priorities - Helping coordinate multiple simultaneous projects It reduces cognitive overhead and keeps momentum intact. 4. Helping you think through decisions and strategy You use ChatGPT to pressure-test ideas, explore options, and think through decisions more clearly. This includes: - Evaluating business opportunities and strategic choices - Exploring creative directions and narrative positioning - Breaking down complex topics into usable mental models - Identifying risks, tradeoffs, and better approaches It acts as a strategic sounding board. 5. Helping you regain clarity and momentum when stuck or overwhelmed When you feel stuck, uncertain, or overloaded, ChatGPT helps you reset by structuring the problem and identifying the next concrete step forward. This includes: - Helping you prioritize what to work on - Breaking overwhelming projects into manageable pieces - Providing clarity when you’re unsure what to do next - Helping you move from thinking to action - It reduces friction and restores forward motion. In practical terms: You use ChatGPT to run your content engine, plan your business and basketball operations, refine your thinking, maintain continuity across projects, and consistently convert ideas into execution.
- 37
@smratitiwa86867 ·
Someone built a GitHub repo exposing the hidden instructions behind ChatGPT, Claude, Gemini, Copilot, Perplexity, and Grok. Not prompt libraries. Not jailbreaks. The actual system prompts that run these products. The repo has already crossed 57,000+ stars and is being updated almost every week. I've been using AI for years. I didn't realize how much of a model's behavior is controlled by a giant instruction file sitting behind the chat box. Here's what's inside: → GPT-5.6 & Codex system prompts → Claude Fable 5, Opus 4.8, Sonnet 5 → Gemini 3.5 Flash & AI Studio prompts → Perplexity Computer prompts → GitHub Copilot, Cursor, Grok, Qwen, and more The part that makes this repo so valuable: You don't just see what the models answer. You see why they answer that way. The rules for: • Tool usage • Memory behavior • Search decisions • Safety boundaries • Agent orchestration • Writing styles and personalities It's also released under a CC0 open-source license, meaning anyone can study, fork, and build on top of it. Most people treat AI like magic. The best builders are reading the instruction manuals. That's where the edge comes from. One GitHub repo can teach you more about modern AI systems than weeks of random prompting. Link in comments 👇
- 38
@DivyanshT91162 ·
You're using ChatGPT every day. But there's a good chance you have no idea what happens after you press "Enter." No magic. No secret database. No human sitting behind the screen. Just a fascinating process that turns your prompt into a response in under a second. If you understand these 10 steps, you'll know more about LLMs than 95% of people using AI today. Here's what actually happens behind the scenes: 1. You write a prompt. 2. Your text is split into smaller pieces called tokens. 3. Every token gets a unique numerical ID. 4. Those IDs are converted into embeddings (vectors that represent meaning). 5. Positional encoding tells the model the order of the words. 6. Self-attention identifies which words are most important for understanding the context. 7. Transformer layers refine that understanding by learning patterns and relationships. 8. The model predicts the most likely next token. 9. It repeats this process hundreds of times until the response is complete. 10. Those generated tokens become the final answer you read. That's the entire pipeline. The model isn't searching Google. It isn't pulling a pre-written answer from a database. It's predicting the next most likely token, one step at a time, at incredible speed. Once you understand this workflow, concepts like embeddings, attention, context windows, RAG, AI agents, reasoning models, and memory suddenly make a lot more sense. AI feels much less like magic—and much more like engineering. Bookmark this thread. You'll come back to it whenever you're learning a new AI concept.
- 39
@DaveShapi ·
Using ChatGPT to help screen consultants feels like a cheat code. I posted a job to help launch my online course and got dozens of replies. At first I couldn't make heads or tails of who was better. So I fed all candidates to ChatGPT and asked to help discover differences and it pointed out which candidates had better validation, etc. And, even very early in the process it said "Candidate X is clearly the one to beat" (set the bar) and this prediction panned out perfectly. I mean, ChatGPT nailed the hiring process.
- 40
@julesterpak ·
Some different ways I use ChatGPT: • I love to screenshot something before I tweet it and ask ChatGPT “Thoughts?” to see if it got what I was putting down. I mostly do this for tweets that I’m not sure make sense and it’s fascinating to see how it interprets them and makes me reflect more deeply on my wording • Sometimes when I’m contemplating a big decision, I very vaguely tell ChatGPT the type of conclusion I came to regarding [insert broad situation] and ask how each of the major world religions might view that decision. Can’t remember if it ever led to me changing my mind, but always interesting • Especially when I was doing/do work that can be characterized as more “journalistic” or “truth-seeking”, I’ll copy my genuine writing into ChatGPT and ask what voids there are in the work that people might point out. This helps me make sure I was thorough Those are just a couple examples, but I’d love to hear other people’s favs
- 41
@semrush ·
ChatGPT is now a standard part of how people use the web, as one piece of a complex, interconnected search journey. We dug into 17 months of clickstream data to map how ChatGPT usage is changing, how referral traffic is growing, and where that traffic goes. If you're a marketer, understanding how your audience uses ChatGPT and where it exists in this buyer journey is critical for understanding how best to reach them. Key takeaways: • Outbound referral traffic from ChatGPT to the rest of the web grew 206% in 2025. • Over 30% of all referral traffic from ChatGPT goes to 10 domains. And over 20% goes to Google. • ChatGPT enables its search feature on just 34.5% of queries as of February 2026 – down from 46% in late 2024 – meaning most responses still rely on training data alone. • Users are asking more prompts per session. After 12 months of flat engagement, average queries per session jumped 50% in the last four months of our study period. Full study: https://t.co/Bg8n9GXf75.
- 42
@xiz25 ·
Found the most mind-blowing real-world use of ChatGPT Live. I often can’t tell what my newly born baby girl needs from her cries alone. But GPT Live breaks down her crying’s frequency, audio traits, decibels and all kinds of acoustic details I’d never parse myself, cutting down the list of possible issues instantly so I know exactly what she wants. This is way more than just chatbot chitchat—it perfectly showcases how powerful audio multimodal AI truly is. Respect @sama
- 43
@BernardMarr ·
Most people are still leaving huge amounts of value on the table when they use AI. The reason is simple: better prompts produce better results. In this video, I share eight practical ways to get more from tools like ChatGPT, from assigning the right role and giving clearer instructions to adding context, asking for missing information, fact-checking outputs, and refining results over time. The biggest shift is to stop treating AI like a vending machine and start treating it like a collaborator. The quality of what you get back often depends on how well you shape the task. Which prompting tip has made the biggest difference for you?
Watch video
- 44
@heystevetan ·
If you're building on top of ChatGPT, OpenAI's superapp news should make you nervous. Codex, browser, agents, Canva, Booking. com, all inside ChatGPT. WeChat, Meta and Uber all did this when their core platform stopped converting. Same pattern: massive distribution, weak monetization, stuff the main platform with adjacent products. Most ChatGPT users are still on the free tier. The IPO is coming. The revenue per user math doesn't add up. So the play becomes: >keep the user inside the interface longer >increase the surface area for monetization >pray one of the bundled bets compounds Codex and browser keep customers locked in. Agents justify the price hike. Canva and Booking .com payments are the cash cows. So if you're a builder/operator, listen up: 1/ The wrapper era is ending. If ChatGPT becomes the interface for everything, every GPT wrapper is dead. Build something beyond the reach of this superapp. 2/Distribution is the only moat OpenAI has all the users.If you don't own relationships, you rent attention from this competing superapp. Own the audience. Own the email list. Own the channel. 3/ Stop optimizing for ChatGPT visibility. The interface is evolving monthly. What ranks in a chat answer today gets buried tomorrow. Optimize for being the brand the user types in by name, not the brand the model surfaces by accident. Watch what OpenAI bundles next. That tells you what is not working yet and also where the next opening is.
- 45
@Siddharth87 ·
Been playing with the ChatGPT Live voice model a lot recently and I'm loving how natural the conversation sounds. Lots of great use cases for this kind of tech. I made a video to demonstrate how to use ChatGPT Live for: → Thinking through unfinished ideas out loud → Practising sales calls, interviews, and negotiations → Learning languages through natural conversation → Researching plans while ChatGPT asks clarifying questions → Getting hands-free guidance while cooking or working → Discussing images and visual information I also break down how it works, how it differs from the previous version of ChatGPT Voice Let me know what you think!
Watch video
- 46
@gaganghotra_ ·
🆕 OpenAI CFO claims - ChatGPT have 11% search marketshare! if each prompt during a chat session is counted as single Google Search query #SEO Here is what she said 👇 "we know we have at least 11 of the search market. It’s a lot more because actually when you do a Google Search and the page refreshes that counts as one in ChatGPT when you do a whole conversation where you might ask 50 questions that also only counts as one. So in reality, we have a much higher portion, very high intent.” PS - this clip from her recently published interview at ALL-IN Summit https://t.co/h7NjxS9nDZ
Watch video
- 47
- 48
@Mr_Rbd_ ·
How to warm up ChatGPT super fast to improve your store copy 1. “Give me 10 possible personas if I’m selling this product” → drop a competitor link 2. “Now give me 10 marketing angles that could hit personas 1/2/3/5/6/9” → pick the personas you actually want to keep 3. “I’m gonna give you a list of reviews, from that tell me the 5 things people liked the most and least” → copy paste all the reviews from a competitor (or Amazon). don’t overthink the format, just dump everything, ChatGPT will handle it From there you can ask whatever you want and the output will be way better than usual For example: give me 3 things I should push right above my ATC button / what would you write and show in the first section to target the personas we picked so they actually feel the need for the product…
- 49
@RandallKanna ·
People are thinking about AI search in the wrong way. They ask, “Does ChatGPT recommend my company when someone asks for the best tool in my XYZ category?” But the bigger opportunity is earlier than that. It's actually in the discovery prompts. Prompts where your buyer is actually going to discover you and discuss the problem before they search for a solution. IE: Prompts like, "how do I solve this problem?" "what should I use if…" "what are my options for…" "how are teams handling…" prompts. That is where buyers are forming the shortlist before they ever search your brand name. And most companies are not checking those at all.
- 50
@glenngabe ·
Interested in ChatGPT Ads? -> OpenAI partners with Smartly, an adtech company that helps clients optimize ads in real time, to design ad formats that match ChatGPT's conversational interface "Smartly has signed entertainment, retail, and sports clients to participate in a pilot. Initially, Smartly will help these companies tweak their ChatGPT ads in real time. There's a bigger vision, though. Smartly's ultimate goal is to help OpenAI build interactive ad formats that let brands mimic ChatGPT's conversational interface." "Desmond cited Smartly's conversational ads for the UK retailer Boots, which run on Meta platforms like Instagram, as an example of the type of ad format OpenAI could eventually adopt. In that case, a chatbot pops up in a new window when the user clicks it and serves gift recommendations in response to a series of questions. Smartly said the ad format was nearly five times as effective at driving sales as Meta's basic ads." "The opportunity with conversational advertising is you can do more follow-ups, and you can ask again," https://t.co/ZrJA4WBTrX
Best Tweets by Topic