Artificial intelligence just did something that nobody expected. In only 48 hours, the global software industry lost over one trillion dollars in market value. The reason? Goldman Sachs started using Claude AI to replace human accountants and back-office workers.
This is not a drill. This is the real thing.
Last week, the entire software industry woke up to a nightmare. Companies that were worth billions yesterday are fighting for survival today.
On trading screens around the world, the red numbers kept growing. The stocks of software giants were falling like stones.
Salesforce, Adobe, and dozens of other companies that once ruled the tech world are now watching their value disappear. These were the golden children of Wall Street. Now investors are running away.
Over three thousand billion dollars of wealth vanished into thin air. The panic spread from New York to London to Tokyo. There was no warning. There was no time to prepare.
our dream ai
The fear started from a single source. The market suddenly realized that a war had begun, and nobody was ready for it.
Software companies can no longer pretend this is not happening.
Many investors thought this was just another AI hype cycle. They were wrong. Goldman Sachs just proved that Claude AI can ai slut do the work of entire accounting departments.
With one move, the game changed forever.
From now on, many jobs that accountants used to do will be done by AI agents instead.
girlfriend gpt
The market has already voted. In the age of AI, whoever moves fast will win. Whoever waits will die.
Goldman Sachs chose to stand on the winning side.
But before you panic, let us look at what is really happening and how you can protect yourself.
Goldman Sachs Uses AI to Replace Thousands of Workers
Just recently, Goldman Sachs chief information officer revealed something shocking. The bank has been working with Anthropic for months to build AI agents that can handle back-office tasks.
These AI agents are already working inside Goldman Sachs. They are reading documents, checking numbers, and handling routine tasks that used to require human workers.
Last year, everyone was talking about Devin, the AI software engineer. But what Goldman Sachs is doing with Claude is much bigger. Claude is not just writing code. It is running entire business processes.
And here is the part that scared investors. Goldman did not spend billions on this. They simply plugged Claude into their existing systems.
According to Marco Argenti, Goldman Sachs chief information officer, Anthropic engineers have been working inside Goldman Sachs since June last year. They built AI systems that can handle the boring but important work that keeps the bank running.
Today, these AI agents are already writing code, talking to trading systems, creating reports, and handling client data.
The key reason is simple. Goldman Sachs processes a huge amount of back-office work every day. Now Claude can do most of it automatically.
Thousands of Workers Face AI Replacement
Under the new partnership with Anthropic, Goldman Sachs plans to use thousands of AI agents across the bank. The speed of this change is frightening.
The fear of losing jobs is no longer just talk. It is happening right now at Goldman Sachs.
Chief information officer Marco Argenti made it clear. He said talking about job losses at this point is too late.
He explained that Goldman Sachs has been running AI projects across the company. These projects are not just experiments. They are real systems that are replacing human workers.
He said something even more direct. He said the people who are being replaced are not just junior staff. They are experienced workers too.
Argenti said Goldman Sachs is starting to see real results. The success of these AI projects is not just about saving money. It is about building a new kind of company where AI handles the boring work and humans focus on big decisions.
More importantly, AI is becoming the core of how Goldman Sachs operates.
Think about it. Document review, risk analysis, data entry, and reporting. These are complex and valuable jobs for humans. But for AI, they are easy tasks that can be done in seconds.
Anthropic Kills the Software Industry and OpenAI Fights Back
The software industry is facing an earthquake. But how did Anthropic cause all this damage?
OpenAI is not sitting still. They are fighting back hard.
Recently, OpenAI president Greg Brockman dropped a bombshell announcement.
He said OpenAI engineers are working on a complete rewrite of their coding tools.
Because of this, OpenAI engineers can basically say goodbye to their old jobs.
Just 12 months ago, Codex could only write simple code snippets. It was interesting but not very useful. Now everything has changed.
Today, Codex has taken over most of the coding work inside OpenAI. It writes code, fixes bugs, and handles complex programming tasks.
This is a historic turning point.
Imagine every company in the world standing on the same starting line. The old tech giants no longer have an advantage. The game has been reset.
To win this war, OpenAI has started a full team restructure. They are moving to a new project-based model.
Their goal is extremely ambitious. The deadline is March 31st.
To reach this goal, OpenAI issued an internal war directive. Every engineer must follow new rules.
The rules are simple but strict. Codex is no longer just a helper. It is now the main worker. Codex 5.2 and the upcoming 5.3 are powerful enough to handle entire projects from start to finish.
Each team must assign one Agent Captain. This person is responsible for managing how AI agents do the work.
Starting next week, the whole company will stop for one day to hold a Codex Hackathon. The goal is to find and fix every problem that stops Codex from working perfectly.
They also created something called AGENTS.md. Every project must have this file. When an AI agent opens the project, it can immediately understand what to do.
They also built a Skills system. When engineers teach Codex a new skill, they must write clear instructions and submit them to a shared library. This way, every person in the company can use that skill.
They even built internal tools to make sure every system has the right connection points. CLI, MCP, and other interfaces allow AI agents to connect directly.
The most extreme rule is this. The model generates code, and the model reviews the code. Human engineers are no longer the main workers. They become reviewers who check what the AI produces.
These measures are making the work process stronger and more efficient. Teams need to succeed through observation, not through chasing bugs, not through chasing code coverage, not through chasing AI usage numbers.
There is one more big change. The Agent-first development rule. Every new project must be built with AI agents in mind from day one. All interfaces must be designed for AI, not just for humans.
To put it simply, they are training themselves to think like AI.
Codex is no longer just a coding tool. It is becoming the entire production line.
The Death of Software Companies
Why is OpenAI making such extreme moves? Because they are scared of being killed by Anthropic.
This panic started with a single public post on January 30th.
Anthropic quietly released Claude Cowork in the form of 11 plugins.
Official announcement: https://claude.com/blog/cowork-plugins
In the tech world, this looks like a simple product update. But the market smelled blood.
These plugins allow Claude to directly control software without using the user interface. It can write code, send emails, manage calendars, and handle databases all by itself.
You need to understand something. For the past ten years, the golden rule of SaaS business was simple. The more users a company had, the more money it made. The more features it added, the higher the price. Salesforce built an empire on this idea.
But Anthropic just broke that rule.
With one move called Agent Plugins, Anthropic went straight to the end user and skipped the software layer entirely.
One AI agent can do the work of ten software subscriptions. Why would any business pay for ten different accounts when one AI can do everything?
The AI agent reads data directly from the back end. It creates reports, sends messages, and manages tasks without ever touching the user interface. Traditional software companies are being bypassed completely.
Some people are calling this the SaaSpocalypse. The end of SaaS.
DocuSign is in trouble because AI can now read and sign contracts automatically. Zendesk is in trouble because AI customer service no longer needs a help desk. Stripe is in trouble because AI can handle payments directly.
If AI agents can access the database directly, then the user interface layer that traditional software companies sell becomes worthless.
Top VC Warns the Software Apocalypse Is Real
So what does the end of SaaS really mean? Top venture capital experts are speaking out.
Wall Street Journal reporter Guo Fei and Leonis Capital partner Jay Zhao are two of the most respected voices in the VC world. They shared their honest views on what is coming.
The core fear behind the SaaS apocalypse is this. If AI agents can directly access databases, then why do we need so many SaaS products?
The answer is clear. Many software companies will die. Many shallow products will lose all their value.
Think about it. Those products that are just pretty wrappers around basic functions. Those tools that only provide a nice user interface for simple tasks. They will be killed by AI agents that can do the same work without the interface.
On the other hand, companies that control the core data and key permissions of an industry may become even more powerful. They could become the execution layer that AI agents rely on.
What will happen in the next 5 to 10 years? Jay Zhao said AI agents will replace both software companies and traditional service companies. The line between them will disappear.
In the future, enterprise software systems may no longer exist as we know them. Instead, there will be a layer of AI agents sitting on top of the database, handling everything from data entry to complex analysis.
The value of user interfaces will drop to almost zero. Buttons and menus will become unnecessary. If AI can handle everything through natural language, why would anyone need a complicated software interface?
This is the new flood of the AI age.
History Shows This Has Happened Before
Steven Sinofsky, the former president of Windows who led the PC revolution and the mobile transition, also spoke about this fear. He said many people are scared because they see the old order being destroyed. But he believes this is not a cliff. It is a springboard.
He is right. AI does not care who you are or what you own. The only thing that matters is whether you can use it to create value.
The lesson from history is simple.
We always think we will be the ones who get replaced. But in the end, it is always someone else who gets hurt.
In the past, companies like EMC, Sun, and Lotus fell. But at the same time, new giants like Amazon, AWS, Google, Microsoft, and SAP rose to take their place.
So do not panic.
Yes, some companies will die. Yes, some jobs will disappear. But new companies will be born. New industries will grow.
The opposite is also true. When old businesses and old technologies die, that is often the most exciting moment for new ideas.
Original post: https://x.com/stevesi/status/2019167552794948020
What You Should Do Now
In 1995, when the internet first arrived, most people did not understand it. They thought it was just a toy. They were wrong. SaaS was born.
History proved that those who laughed at the internet were the ones who got left behind.
Ten years later, they were still trying to catch up.
Today, in just 48 hours, the software industry lost over a trillion dollars. But this is not the end. This is the beginning of a new world.
For investors, this is both a painful lesson and a huge opportunity. Before the explosion, nobody believed this could happen so fast. But when the dust settles, those who bought at the bottom may become the biggest winners.
In this war, only those who can adapt will survive. Only those who are brave enough to face the future will thrive.
