geekbing
banner
geekbing.bsky.social
geekbing
@geekbing.bsky.social
Independent Developer | Value Investor
US Earnings Analysis ‣ ch2report.com
Pinned
Financial data websites on the market are often overwhelming.

So I built a minimalist financial statement website.

All key data on one screen.

ch2report.com
When developing a chatbot application, a Markdown rendering component is almost always required.

Unfortunately, Streamdown only supports React.

I finally found a great Vue/Nuxt Markdown rendering library, and its documentation is very detailed.

github.com/jinghaihan/v...
January 15, 2026 at 11:06 AM
Meta’s gross margin is genuinely impressive. It has barely dipped below 80% over the past decade, so a $2 billion acquisition of Manus is basically pocket change for them.
January 7, 2026 at 5:07 AM
Many people don’t know how to analyze a public company’s financial statements.

In fact, you only need to focus on the three core statements: the income statement, the balance sheet, and the cash flow statement.
January 4, 2026 at 12:35 PM
With vibe coding, it’s really easy to build all kinds of little toys.

But when it comes to complex production apps, you still have to put in a lot of work.

All the engineering issues can be pretty overwhelming 😆
December 27, 2025 at 11:56 AM
🎉 A new version of Chat2Report is out!

✨ New: Instant chart-based trend analysis for U.S. stocks, covering the latest 10 years and 20 quarters of financial statement metrics 📈

Check out the full details in a 1MB video 🎬

Try it now 👉 ch2report.com
December 2, 2025 at 1:59 PM
Struggling with open-source code or new components?

Try DeepWiki. It turns GitHub repos into clear docs with charts, and you can chat with the code like ChatGPT.

deepwiki.com
December 1, 2025 at 2:46 AM
The Vue community has introduced the AI Elements Vue project.

It closely mirrors Vercel’s AI Elements and is built on top of the Shadcn system.

A modern component library that makes it easy to build AI applications.

Hope more Vue developers get to know it.

github.com/vuepont/ai-e...
GitHub - vuepont/ai-elements-vue: AI Elements Vue is a component library built on top of shadcn-vue to help you build AI-native applications faster.
AI Elements Vue is a component library built on top of shadcn-vue to help you build AI-native applications faster. - vuepont/ai-elements-vue
github.com
November 26, 2025 at 10:52 AM
This password manager is pretty solid and works on both Windows and macOS.

The desktop version is free, and your data can sync through Google Drive, Dropbox, OneDrive, iCloud and more.

The mobile version requires a one-time purchase for lifetime use.

safe-in-cloud.com/en/
November 22, 2025 at 3:59 AM
🎉 Chat2Report new version is live.

US stock financials with the Big Three Statements and Ratio Data(Last 10 years and 20 quarters) ✨

Super easy to use. Just watch the demo 🎬

Coming:
F-Score rankings, S&P 500, Nasdaq-100

Tap any row in the statements or ratios and it instantly becomes a chart 📈
November 21, 2025 at 11:51 AM
I've been switching between GPT-5.1 Codex High, Codex Max High, Gemini 3 Pro, and GLM-4.6.

Now I have a setup that works.

Front-end across files: Codex Max High or Gemini 3 Pro.

Complex back-end: Codex Max High.

Simple tasks: GLM-4.6.
November 20, 2025 at 5:22 AM
No sooner had Gemini 3 Pro been released than GPT-5.1 Codex Max followed hot on its heels.

Competition in the AI world is now measured in days.

openai.com/index/gpt-5-...
November 20, 2025 at 2:00 AM
A few days ago people were mocking Buffett for buying Google. Now it completely backfired.

The Oracle of Omaha always proves why he deserves trust.
November 19, 2025 at 2:51 PM
Gemini 3 Pro is indeed powerful, resolving front-end issues across multiple files.

It outperforms GPT-5.1 Codex High in speed while consuming fewer tokens.
November 19, 2025 at 11:31 AM
Same model GPT-5.1 Codex High

Cursor's Agent mode is actually the strongest

Codex is just a little brother in comparison

GLM-4.6 in Claude Code gets completely destroyed
November 19, 2025 at 3:49 AM
Robyn is quite an impressive Python web framework.

Requests per second (RPS) comparison:
🦀 Robyn ████████████████ 100%

⚡️ FastAPI ██████████ 60-70%

🏃 Sanic ████████ 50-60%

🌶️ Flask ███ 15%

🟢 Django ██ 10%

robyn.tech/documentatio...
Moreover, these tests were conducted on my development machine, and thus, the figures portrayed below are not absolute. The numbers only serve to indicate the relative performance of these frameworks under the specific testing conditions.
robyn.tech
November 19, 2025 at 2:37 AM
The newly announced Gemini 3 pricing is honestly laughable.

Gemini 3 Pro (Preview):
≤ 200K tokens
Input: $2
Output: $12

>200K tokens
Input: $4
Output: $18

GPT-5.1:
1M tokens
Input: $1.25
Output: $10

Who approved this pricing?

There’s absolutely no competitive advantage.
November 18, 2025 at 4:31 PM
Cloudflare's service is down, causing a bunch of websites to go offline. Hurry up and buy CF stock.

#Stock
November 18, 2025 at 1:30 PM
Grok 4.1 is really low-key.

Its creative writing is second only to GPT-5.1.

Without the overwhelming hype some models push.

x.ai/news/grok-4-1
Grok 4.1 | xAI
Grok 4.1 is now available to all users on grok.com, 𝕏, and the iOS and Android apps. It is rolling out immediately in Auto mode and can be selected explicitly as “Grok 4.1” in the model picker.
x.ai
November 18, 2025 at 3:02 AM
Traction: How Any Startup Can Achieve Explosive Customer Growth" is a fantastic read.

Every indie developer and content creator should check it out. Myself included.
November 16, 2025 at 10:55 AM
Front-end pages designed by AI often look the same nowadays.

Try exploring the ideas below to make your designs more unique.

claude.com/blog/improvi...
November 16, 2025 at 2:45 AM
This resource monitoring terminal is more fun than htop.

I highly recommend it. It’s called btop.

github.com/aristocratos...
November 15, 2025 at 2:17 AM
Using GPT-5 Codex High is still the best choice for handling complex tasks, especially those involving multiple files.

For simpler tasks, GLM 4.6 offers unbeatable cost performance, and the GLM Coding Lite plan provides 120 prompts every 5 hours.
November 15, 2025 at 2:02 AM
After trying out different models, including Claude, Gemini, Grok, and ChatGPT, I still find ChatGPT's web search to be the most useful, with the most accurate results.

#AI
November 15, 2025 at 1:30 AM
Gabe, looks like you got stuck right before hitting three again, huh?
November 13, 2025 at 2:10 PM
Which is more suitable for the financial sector?

GPT, Claude, Gemini, or Grok?

#AI #Finance #Stock
November 13, 2025 at 12:56 PM