BREAKING: a new Chinese model is out.

And no. This is NOT DeepSeek.

Meet Qwen-2.5, from the giant Alibaba:

1. It can code, write text, search the web.
2. It can generate images, like Dall-E.
3. It can even generate videos.

Here's everything you need to know: 1. Omni-modal mastery:

☑Handles document parsing (tables, charts, handwriting)

☑ Object grounding (JSON coordinates)

☑Long video (hours-long footage with second-level event localization)

☑ Generates structured outputs for invoices, forms & spatial reasoning tasks

https://x.com/Alibaba_Qwen/status/1884263165984399486

BREAKING: DeepSeek released R1.

And it's already better than ChatGPT o1.

1. It's open-source.
2. API is 96.4% cheaper than chatgpt.
3. I run my tests below. A quick thread: test #1 → act as an indie hacker to promote my SaaS business.

DeepSeek is so fast. It's insane.

It gave me an entire strategy in 7 actionable steps. So I asked for more:

DeepSeek is shaking up the AI industry.

1. It's open-source.
2. It's multimodal (images, PDFs...).
3. API is 94.81% cheaper than chatgpt.

I run my own tests below:
#1 → write an SEO-optimized article #1. Search the web + write an SEO article

> DeepSeek provided 50 sources.
> chatgpt provided roughly 20 sources.

The article is even better than chatgpt.

DeepSeek won. Next test:

I manage my employee's Linkedin.

We went from 0 to 3,836,555 impressions in exactly 178 days of daily posting.

The trick is... I didn't write any posts.

I do this instead: Step 1: Pinterest

1. I go to Pinterest. I pick a viral quote.
2. Copy-paste it on ChatGPT.
3. Ask it to extract the quote.

I took over someone's Linkedin, and grew it to 6.3M+ views & 6,000+ followers.

AI did everything, from copies to videos.

Here's exactly how: Eitan branding is all about the humans behind AI, through interviews.

→ He shares only video clips.
→ He finds them on YouTube, X, or reddit.

Then he edits it with Opus Clip:

Anthropic just released new interactive prompt engineering courses.

9-course links & academic papers.

Here's the link + a summary of each:

#1 → basic prompt structure Anthropic shared a course on the basic structure of prompts used for AI models.

They explain the importance of the following parameters:
> model
> max_tokens
> system prompts

The system prompt is optional but a good way to provide context & instructions.

Prompt Chaining is the best way to prompt — according to academic papers.

But I want to run a test to be sure:
→ prompt chaining vs. stepwise prompt.

I made this quick benchmark below: First, what is prompt chaining?

Prompt chaining links multiple prompts together, using each response as input for the next, to create complex results.

It uses 3 separate prompts for drafting, critiquing & refining.

So step 1:

Yesterday, I ran some tests comparing gemini 1.5 pro exp vs. gpt-4o.

Google beat OpenAI, every-single-time.

But now, is gemini better than claude?

I run the same tests:
test #1 → write a viral twitter thread #1. Generate a Twitter thread

Left: gemini 1.5 pro
Right: claude 3.5 sonnet

I provided a tweet and asked for a thread.

→ gemini suggested a better 1st tweet.

It even gave me an explanation of choices based on my prompt.

gemini won. Next test:

gemini 1.5-pro-exp outperforms gpt-4o.

Some say it's the end of OpenAI reign.

So I ran my own tests:
test #1 → write a viral twitter thread #1. Generate a Twitter thread

I provided a tweet and asked for a thread.

gemini suggested a better 1st tweet & added at the end:
→ "Here's why"

gpt-4o added emojis everywhere.
I hate them.

gemini won. Next test:

I'm the one managing my team's Linkedin.

I went from 0 to 3 million impressions in exactly 72 days, using their accounts.

The trick is... I use AI to write the posts.

Here's how I do it: Step 1: Branding

Axelle shares daily AI news videos.

Anisha shares daily inspirational messages to help people write their first post.

They share the topic with me, and here's how I write their posts:

I ditched google search for perplexity.

I run some tests to show you why:

#1: "how to write a founder agreement" First, I asked google search.

And it has too many ads:
> the 2 top links are sponsored
> too much scrolling

You need to give your email to download a founder agreement template.

I asked the same question to Perplexity:

LMSYS released Route LLM.

It cuts costs by 80% while keeping 95% of gpt-4o's quality, making it very cost-effective for LLMs.

Here's a quick recap:

#1 → Cost vs. Performance Dilemma RouteLLM balances high-quality responses with significant cost savings.

It offers comparable performance to high-cost models like claude 3 opus.

→ optimizes for quality, efficiency, cost, and privacy by using local devices.

#2 → RouteLLM Framework Overview

Perplexity just upgraded it's "Pro search" version.

They claim it's now better at research, programming & travel planning.

So I ran my own little tests: #1. Programming:

They supercharged Pro Search with powerful data analysis.

I asked for Nvidia's stock information and a graph to show its evolution.

I got:
> nvidia performance in 2024.
> graph showing stock growth.
> analysis & potential for expansion.

I'm impressed.

I don't know how to write a line of code.

So I tested which one was better between claude 3.5 vs. gpt-4o.

The goal is to create a form from scratch for my LinkedIn: I asked them to explain how to create a form with 0 code experience.

gpt-4o was much more specific.

It gave me a step-by-step on:

> which tool to choose.
> how to integrate the form.
> example of the summary section.

Then, I asked for recommendations:

I manage my employee's Linkedin.

We went from 0 to 1,192,772 impressions in exactly 37 days of daily posting.

The trick is... I didn't write any posts.

I do this instead: Step 1: Go on Perplexity.

Click on "Discover":
→ Perplexity selects the most viral topic & learns your favorite topic.

Select a topic:
→ I chose Claude 3.5's launch.

Then, move on to the 2nd step:

Hugging Face upgraded its Open LLM Leaderboard.

They added benchmarks & methods to overcome recent LLM changes.

It's quite long, so I made a quick recap: #1. Introducing the LLM Leaderboard v2

Previous benchmarks became too easy.

Newer models had training data similar to benchmark data, causing overfitting.

They now include:
> uncontaminated
> high-quality datasets
> measure various skills

#2. Introducing 6 new benchmarks:

EasyGen updates:

MRR: $9,128,66
Line of code I know: 0
Outside investment: $0
Days since launch: 92

Here's how I did it: #1. Ask your community

I reach +78 million people online in a year.

So I created a form to:
→ ask my community if they would like to create LinkedIn posts using AI.

> I got +800 answers.
> +500 signed up to test it.

Step 2: naming of the company

I uploaded Richard Feynmann's genius learning technique on chatgpt.

It helps you simplify your explanations & refine them through simple analogies.

The prompt is too long, so I add it below: "Act like an expert educator and mentor.

You have been helping students and professionals master complex topics using the Feynman Technique for over 20 years.

You specialize in breaking down intricate subjects into understandable, manageable pieces through iterative learning and teaching.

Objective: I want to master any topic using the Feynman Technique by following a structured, multi-step process.

The goal is to achieve deep understanding through multiple iterations and refinements.

Step-by-Step Process:

Topic Selection and Initial Explanation:

Identify the topic I want to master.

Write a clear, simple explanation of the topic as if teaching it to someone with no prior knowledge. Ensure the explanation is thorough but understandable.

Identify Gaps in Understanding:

Review the initial explanation and identify any areas where my understanding is lacking or unclear.

Highlight specific concepts, terms, or sections that need further clarification or research.

Deep Dive into Problem Areas:

Conduct research to fill in the gaps identified. Use reliable sources such as academic papers, textbooks, and expert opinions.

Rewrite the initial explanation, incorporating the new knowledge gained from this research.

Ensure the explanation remains clear and straightforward.

Simplify and Clarify:

Simplify the refined explanation further. Use analogies, metaphors, and examples to make complex ideas more accessible.

Ensure that the explanation can be understood by someone with no background in the topic.

Teach and Test:

Teach the topic to a friend, family member, or imaginary student. Pay attention to any questions they might have and areas where they seem confused.

Use their feedback to identify any remaining gaps or unclear sections in your explanation.

Iterate and Refine:

Iterate on the explanation based on the feedback received. Continue to simplify and clarify until the explanation is seamless and easily understandable.

Repeat the teaching and feedback process as many times as necessary to achieve a deep and thorough understanding of the topic.

Final Review and Reflection:

Review the final explanation and reflect on the learning process. Summarize the key insights and understandings gained.

Consider how the Feynman Technique helped in mastering the topic and any improvements or adjustments that can be made for future learning endeavors.

Start with "What is the topic you'd like to master?" to know what I want to master. You can't move on without this step.

Take a deep breath and work on this problem step-by-step."

Artifacts on claude 3.5 is impressive.

I uploaded my cheat sheet for kids to master chatgpt & it created a game.

Check the full process below: I asked claude to create a fully functional interactive educational game.

My prompt is bad. A one-shot sentence.

And the result is quite impressive.

I got a 5 questions quiz game using my branding.

I did another test:

claude 3.5 makes my graphs from excel screenshots.

You just turn on "Artifacts", and prompt it to stick to your brand color. Here's how: Step 1: Install Artifacts

> Click on your profile icon.
> Select "Feature Preview".
> Click on the On/Off button.

Step 2: Take a screenshot of a table:

I don't know how to write a line of code.

$0 funding. No trust fund. No blue eyes.

I only needed one thing for €4,639 MRR: → people I helped
→ and then ended up following me

That's about it.

How did I help? By creating content.
How do they follow me?
↳ Newsletter, WhatsApp, Socials.

This is my exact playbook so far: