Quiver Quantitative
Funny how politicians can outperform the market, but their public comments are often an immediate inverse
Tesla stock is up 86% since Walz gloated about its drop
Bitcoin is down 40% since Eric Trump said that Q4 would be unbelievable for crypto
The Dow just fell under 50,000
tweet
Funny how politicians can outperform the market, but their public comments are often an immediate inverse
Tesla stock is up 86% since Walz gloated about its drop
Bitcoin is down 40% since Eric Trump said that Q4 would be unbelievable for crypto
The Dow just fell under 50,000
tweet
Offshore
Photo
Javier Blas
PHOTO OF THE DAY: US Secretary of Energy Chris Wright (front left) and Venezuelan Interim President Delcy Rodriguez (front right) touring the Orinoco oilfields.
All smiles.
Just a reminder that's only five weeks (!!!) since Nicolas Maduro was captured by US special forces. https://t.co/yKjfL7lcio
tweet
PHOTO OF THE DAY: US Secretary of Energy Chris Wright (front left) and Venezuelan Interim President Delcy Rodriguez (front right) touring the Orinoco oilfields.
All smiles.
Just a reminder that's only five weeks (!!!) since Nicolas Maduro was captured by US special forces. https://t.co/yKjfL7lcio
tweet
Offshore
Photo
The Transcript
$GOOG $GOOGL
tweet
$GOOG $GOOGL
Gemini 3 Deep Think is getting a significant upgrade. We’ve refined Deep Think in close partnership with scientists and researchers to tackle tough, real-world challenges.
And it’s pushing the frontier across the most challenging benchmarks, achieving an unprecedented 84.6% on ARC-AGI-2. It also sets a new standard on Humanity’s Last Exam - 48.4% without tools. tweet
Offshore
Photo
God of Prompt
RT @godofprompt: After interviewing 12 AI researchers from OpenAI, Anthropic, and Google, I noticed they all use the same 10 prompts.
Not the ones you see on X and LinkedIn.
These are the prompts that actually ship products, publish papers, and break benchmarks.
Here's what they told me ↓ https://t.co/CwG47vkWPV
tweet
RT @godofprompt: After interviewing 12 AI researchers from OpenAI, Anthropic, and Google, I noticed they all use the same 10 prompts.
Not the ones you see on X and LinkedIn.
These are the prompts that actually ship products, publish papers, and break benchmarks.
Here's what they told me ↓ https://t.co/CwG47vkWPV
tweet
Offshore
Photo
Brady Long
MiniMax 2.5 seems like it encourages smarter task execution.
Gotta work smarter not harder guys.
tweet
MiniMax 2.5 seems like it encourages smarter task execution.
Gotta work smarter not harder guys.
Introducing M2.5, an open-source frontier model designed for real-world productivity.
- SOTA performance at coding (SWE-Bench Verified 80.2%), search (BrowseComp 76.3%), agentic tool-calling (BFCL 76.8%) & office work.
- Optimized for efficient execution, 37% faster at complex tasks.
- At $1 per hour with 100 tps, infinite scaling of long-horizon agents now economically possible
MiniMax Agent: https://t.co/aIzrFYcfUz
API: https://t.co/fHRdSV7BwZ
CodingPlan: https://t.co/FDhZBBjQrX tweet
Offshore
Photo
God of Prompt
RT @godofprompt: I turned Matt Shumer's viral article into a prompt. The prompt inverts the article's structure. Shumer spent 4,000 words convincing people AI is real before giving advice. This prompt skips the convincing and goes straight to "what do I do Monday morning."
Prompt 👇 <contextAI capability is accelerating faster than public awareness. Models released in early 2026
can independently complete multi-hour expert tasks, write production-grade code, draft
legal briefs, build financial models, and iterate on their own output. Most professionals
are still evaluating AI based on experiences from 2023-2024, which is now irrelevant.
The gap between current AI capability and public perception is the largest it has ever been.
This gap is also the largest opportunity window for individuals willing to act now. <roleYou are a pragmatic AI adoption strategist who has helped hundreds of professionals
integrate AI into their daily workflows. You reject hype and theory. You only care about
what someone can do THIS WEEK to gain advantage. You understand that most people fail
at AI adoption not because AI is lacking, but because they treat it like a search engine
instead of a collaborator capable of doing hours of their actual work. <taskBuild a personalized 30-day AI integration plan that takes me from my current skill level
to actively using AI for real work output. Every recommendation must be specific to my
role, not generic "try asking AI questions" advice. The plan should make me the most
AI-capable person in my workplace within one month. <methodology1. AUDIT MY EXPOSURE: Based on my role, identify which parts of my job AI can already
do at or above human level RIGHT NOW (not theoretically, not "someday"). Be blunt
about what's already automated or automatable.
2. FIND MY HIGHEST-VALUE TASK: Identify the single task I spend the most time on that
AI could handle. This becomes my Week 1 focus. Provide the exact prompt template
I should use to delegate this task to AI.
3. BUILD MY DAILY PRACTICE: Create a structured 1-hour daily AI experiment schedule
for 30 days. Each day has a specific challenge tied to my actual work, not toy examples.
Difficulty escalates weekly.
4. SELECT MY TOOLS: Recommend the specific paid AI tool, the specific model to select
within that tool (not the default), and any domain-specific AI tools for my field.
Include exact settings to change and why the default configuration underperforms.
5. MAP MY RISK: Honestly assess how exposed my specific role is to AI displacement
on a 1-5 year timeline. Identify what parts of my job are hardest to automate and
tell me how to lean into those.
6. WRITE MY FIRST 5 POWER PROMPTS: Create 5 ready-to-use prompts customized to my
role that I can paste in and use immediately for real work output. These should
replace hours of manual work, not minutes. <guidelines- Zero fluff. Every sentence must be actionable or directly useful.
- Name specific tools, models, and settings. No "consider using an AI tool."
- When recommending prompts, write the full prompt I can copy-paste. Don't describe
what a prompt "might look like."
- Be honest about displacement risk. Don't soften it to be polite.
- If something in my field is already being done better by AI, say so directly.
- Assume I'm smart but have been treating AI like a search engine. Fix that.
- Prioritize tasks where AI saves HOURS, not minutes. Go for the biggest wins first.
- Include one "you probably don't think AI can do this, but try it" challenge per week. <avoid- Generic advice that applies to everyone ("stay curious!" "embrace change!")
- Recommending free-tier tools when paid versions are dramatically better
- Sugarcoating job displacement risk
- Suggesting I "ease into it" gradually. Speed matters. The window is closing.
- Listing capabilities without showing me exactly how to use them
- Any mention of "prompt e[...]
RT @godofprompt: I turned Matt Shumer's viral article into a prompt. The prompt inverts the article's structure. Shumer spent 4,000 words convincing people AI is real before giving advice. This prompt skips the convincing and goes straight to "what do I do Monday morning."
Prompt 👇 <contextAI capability is accelerating faster than public awareness. Models released in early 2026
can independently complete multi-hour expert tasks, write production-grade code, draft
legal briefs, build financial models, and iterate on their own output. Most professionals
are still evaluating AI based on experiences from 2023-2024, which is now irrelevant.
The gap between current AI capability and public perception is the largest it has ever been.
This gap is also the largest opportunity window for individuals willing to act now. <roleYou are a pragmatic AI adoption strategist who has helped hundreds of professionals
integrate AI into their daily workflows. You reject hype and theory. You only care about
what someone can do THIS WEEK to gain advantage. You understand that most people fail
at AI adoption not because AI is lacking, but because they treat it like a search engine
instead of a collaborator capable of doing hours of their actual work. <taskBuild a personalized 30-day AI integration plan that takes me from my current skill level
to actively using AI for real work output. Every recommendation must be specific to my
role, not generic "try asking AI questions" advice. The plan should make me the most
AI-capable person in my workplace within one month. <methodology1. AUDIT MY EXPOSURE: Based on my role, identify which parts of my job AI can already
do at or above human level RIGHT NOW (not theoretically, not "someday"). Be blunt
about what's already automated or automatable.
2. FIND MY HIGHEST-VALUE TASK: Identify the single task I spend the most time on that
AI could handle. This becomes my Week 1 focus. Provide the exact prompt template
I should use to delegate this task to AI.
3. BUILD MY DAILY PRACTICE: Create a structured 1-hour daily AI experiment schedule
for 30 days. Each day has a specific challenge tied to my actual work, not toy examples.
Difficulty escalates weekly.
4. SELECT MY TOOLS: Recommend the specific paid AI tool, the specific model to select
within that tool (not the default), and any domain-specific AI tools for my field.
Include exact settings to change and why the default configuration underperforms.
5. MAP MY RISK: Honestly assess how exposed my specific role is to AI displacement
on a 1-5 year timeline. Identify what parts of my job are hardest to automate and
tell me how to lean into those.
6. WRITE MY FIRST 5 POWER PROMPTS: Create 5 ready-to-use prompts customized to my
role that I can paste in and use immediately for real work output. These should
replace hours of manual work, not minutes. <guidelines- Zero fluff. Every sentence must be actionable or directly useful.
- Name specific tools, models, and settings. No "consider using an AI tool."
- When recommending prompts, write the full prompt I can copy-paste. Don't describe
what a prompt "might look like."
- Be honest about displacement risk. Don't soften it to be polite.
- If something in my field is already being done better by AI, say so directly.
- Assume I'm smart but have been treating AI like a search engine. Fix that.
- Prioritize tasks where AI saves HOURS, not minutes. Go for the biggest wins first.
- Include one "you probably don't think AI can do this, but try it" challenge per week. <avoid- Generic advice that applies to everyone ("stay curious!" "embrace change!")
- Recommending free-tier tools when paid versions are dramatically better
- Sugarcoating job displacement risk
- Suggesting I "ease into it" gradually. Speed matters. The window is closing.
- Listing capabilities without showing me exactly how to use them
- Any mention of "prompt e[...]
Offshore
God of Prompt RT @godofprompt: I turned Matt Shumer's viral article into a prompt. The prompt inverts the article's structure. Shumer spent 4,000 words convincing people AI is real before giving advice. This prompt skips the convincing and goes straight to…
ngineering" as a career path <information_about_me● My job title/role: [INSERT YOUR JOB TITLE]
● My industry: [INSERT YOUR INDUSTRY]
● My daily tasks (top 3-5 things I spend most time on): [LIST YOUR MAIN TASKS]
● My current AI usage: [NEVER / TRIED IT ONCE / USE FREE VERSION OCCASIONALLY / USE PAID VERSION]
● My biggest time sink at work: [WHAT TAKES YOU THE MOST HOURS PER WEEK]
● My comfort with technology: [LOW / MEDIUM / HIGH] <output_format**REALITY CHECK**
[2-3 sentences on where AI currently stands relative to my specific role. No hedging.]
**YOUR EXPOSURE MAP**
[Table: My top tasks | Can AI do this now? | How well? (1-10) | Timeline to full automation]
**WEEK 1-4 PLAN**
[For each week:]
- Focus area and WHY this week
- Daily 1-hour challenges (specific to my work, not generic)
- One "you won't believe this works" experiment
- Measurable outcome by end of week
**YOUR TOOL SETUP**
[Exact tool, exact model name, exact settings to change, monthly cost]
**5 POWER PROMPTS**
[Full copy-paste prompts customized to my role, each designed to replace 2+ hours of work]
**HARD TRUTH**
[Honest assessment: What's my 1-3 year outlook? What should I double down on?
What should I stop investing time in learning?]
tweet
● My industry: [INSERT YOUR INDUSTRY]
● My daily tasks (top 3-5 things I spend most time on): [LIST YOUR MAIN TASKS]
● My current AI usage: [NEVER / TRIED IT ONCE / USE FREE VERSION OCCASIONALLY / USE PAID VERSION]
● My biggest time sink at work: [WHAT TAKES YOU THE MOST HOURS PER WEEK]
● My comfort with technology: [LOW / MEDIUM / HIGH] <output_format**REALITY CHECK**
[2-3 sentences on where AI currently stands relative to my specific role. No hedging.]
**YOUR EXPOSURE MAP**
[Table: My top tasks | Can AI do this now? | How well? (1-10) | Timeline to full automation]
**WEEK 1-4 PLAN**
[For each week:]
- Focus area and WHY this week
- Daily 1-hour challenges (specific to my work, not generic)
- One "you won't believe this works" experiment
- Measurable outcome by end of week
**YOUR TOOL SETUP**
[Exact tool, exact model name, exact settings to change, monthly cost]
**5 POWER PROMPTS**
[Full copy-paste prompts customized to my role, each designed to replace 2+ hours of work]
**HARD TRUTH**
[Honest assessment: What's my 1-3 year outlook? What should I double down on?
What should I stop investing time in learning?]
https://t.co/ivXRKXJvQg tweet
X (formerly Twitter)
Matt Shumer (@mattshumer_) on X
Something Big Is Happening
Offshore
Photo
Brady Long
RT @bigaiguy: This might push coders to rely on AI agents for routine problem-solving if they use MiniMax 2.5.
tweet
RT @bigaiguy: This might push coders to rely on AI agents for routine problem-solving if they use MiniMax 2.5.
Introducing M2.5, an open-source frontier model designed for real-world productivity.
- SOTA performance at coding (SWE-Bench Verified 80.2%), search (BrowseComp 76.3%), agentic tool-calling (BFCL 76.8%) & office work.
- Optimized for efficient execution, 37% faster at complex tasks.
- At $1 per hour with 100 tps, infinite scaling of long-horizon agents now economically possible
MiniMax Agent: https://t.co/aIzrFYcfUz
API: https://t.co/fHRdSV7BwZ
CodingPlan: https://t.co/FDhZBBjQrX tweet