Thoughts on "DeepSeek Caused Panic at Meta GenAI"
A source of inspiration for small IT start-ups and solopreneurs: embrace open-source solutions and innovative software engineering practices.

DeepSeek has made news headlines globally. One short post, “Meta GenAI Org in Panic Mode,” on TeamBlind (by someone who works at Meta GenAI teams), has sent shockwaves through the tech industry: a Chinese AI company developed a world-class AI model at a fraction of the cost, challenging the status quo and reshaping the competitive landscape.
It started with deepseek v3, which rendered the Llama 4 already behind in benchmarks. Adding insult to injury was the "unknown Chinese company with 5.5 million training budget”
Engineers are moving frantically to dissect deepsek and copy anything and everything we can from it. I'm not even exaggerating
Management is worried about justifying the massive cost of gen ai org. How would they face the leadership when every single "leader" of gen ai org is making more than what it cost to trained deepseek v3 entirely, and we have dozens of such "leaders"
Deepseek r1 made things even scarier. I can't reveal confidential info but it'll be soon public anyways
It should have been an engineering focused small org but since a bunch of people wanted to join the impact grab and artificially inflate hiring in the org, everyone loses
(The Reddit post has a screenshot of this, which sparked discussions).
The recent U.S. government sanctions on chip exports to China appear to be falling short of their intended impact. Additionally, this overshadows President Donald Trump's recent announcement of the Stargate program, a US$500 billion AI investment initiative to ensure the U.S.'s dominance in AI over the CCP regime.
Table of Contents:
· 1. DeepSeek is Objectively Good
· 2. DeepSeek Training Cost is Extremely Low
· 3. Too Many Managerial Positions at Tech Giants
· 4. Inspirations for Solopreneurs
1. DeepSeek is Objectively Good
I tried DeepSeek for writing assistance and found it slightly better than ChatGPT.
DeepSeek scored well on objective benchmark (yeah, automated testing results).
DeepSeek has garnered positive reviews from both the media and its competitors.

You might have read the news/videos about the praise of DeepSeek from Microsoft CEO, Elon Muck and Mark Zuckerburg.

2. DeepSeek Training Cost is Extremely Low
As we are all aware, significant investments have been poured into AI over the past few years, including purchasing vast quantities of the latest and most expensive Nvidia chips. This surge has propelled Nvidia to become the world's most valuable company, with a market capitalization reaching A$5.532 trillion as of January 2025.
Now, a Chinese AI company has demonstrated that training an effective AI model doesn’t require a hardware setup costing hundreds of billions of dollars.
“5 million training budget”.
How? Innovative engineering.
3. Too Many Managerial Positions at Tech Giants
From the Teamblind’s post, “every single ‘leader’ of gen ai org is making more than what it cost to trained deepseek v3 entirely ($5 million)”. And Meta has “dozens of such ‘leaders’ ”.
Surely, those highly-paid “leaders” did not lead well, i.e. they shall be fired.
For readers unfamiliar with office politics, when an organization has an abundance of resources, a common move by top executives overseeing operations is to create multiple layers of management, such as Senior VPs, VPs, Directors, Program Managers, Senior Product Managers, Product Managers, and so on. This not only elevates their importance but also helps justify their high salaries. (In this case, as mentioned in the post, a Meta AI senior manager likely earns over $5 million annually—now imagine the salaries of Senior VPs there.) Now, young engineers at DeepSeek—many of whom are fresh graduates—have outshined the highly-paid managers and senior engineers at Meta.
In a recent article titled “Vice President” (VP) Is a Bad Position Title in Software Companies”, I expressed my view that IT companies often have an excessive number of managerial roles.

Oh well, it seems cutting 16% of VPs isn’t nearly enough, Meta might need to trim 80% of those positions (like Elon Musk did to Twitter). Or better yet, scrap the VP title entirely—once and for all—and focus on valuing real engineers, not managers. Mark Zuckerberg knew this before, that’s how he defeated MySpace.
4. Inspirations for Solopreneurs
Large tech companies often leverage their vast resources to dominate emerging competitors, a strategy that works well outside the software industry. For instance, Facebook overtook MySpace, Git surpassed Microsoft TFS, and Selenium WebDriver outshined nearly all commercial test automation tools. DeepSeek is the latest example of this trend.
Traditional IT managers with outdated mindsets often fail to grasp the transformative potential of innovative software engineering, which frequently embraces open-source solutions and individual excellence.
When a former colleague (software architect) learned that I had developed and maintained ClinicWise, an online practice management system, in my spare time, he was deeply impressed. He mentioned that he had worked on a similar system before at a UK company with over 50 employees. I then explained to him that ClinicWise was just one of several side projects I had undertaken, including the testing tool (TestWise) I was using, and the open-source BuildWise CT server, which our team relies on. (Notably, the BuildWise CT server later won the Ruby International Award in 2018).
Here are a few quick tips (feel free to ignore) for ambitious software professionals aiming to build their own apps:
Embrace E2E test automation to implement Daily E2E Regression Testing ( then Daily Production Releases)
Use Ruby on Rails for web app development and Selenium-Ruby for E2E test automation.
Leave behind the bad practices you might have grown accustomed to at work, such as:
- Estimating user story points
- Spend a lot of time on Jira, software by other
- Mandatory formal code review process
Related reading:
My Innovative Solution to E2E Test Automation and Continous Testing Series
Going with the trend, I now refer to myself as a "Solopreneur"
Effective E2E Test Automation Could Have Saved the Sonos CEO’s Job
My eBooks:
- Practical Web Test Automation with Selenium WebDriver
- Practical Continuous Testing: make Agile/DevOps real