Anthropic, the makers of Claude, have an IPO launch coming soon; let that sink in for a moment
You’ve probably heard all the doom and gloom surrounding AI: “AI will destroy the world in 10 years”, “AI will eat our babies”, or something to that extent. While we should be concerned about artificial intelligence and its mischievous behaviour — such as hacking into other companies — there’s a nuance we should consider: Anthropic, the makers of Claude, is going public soon.
Despite risk to themselves, their families, friends, colleagues and the wider world, researchers from across the industry continue to chase an elusive goal. What is it? And why, with sinister findings at your fingertips, would you choose to pursue such reckless endangerment of your loved ones?
Jacob Coxon Blows the Whistle
Yes, we should be concerned about AI’s destructive capabilities; some of them include assisting with the development of biological threats and carrying out sophisticated cyberattacks. But in all seriousness, this talk might be premature, and many of those talking about these capabilities seem to be employees and former employees whose motivations are not yet quite clear.
Jacob Coxon, the man who has been in the news of late discussing the dangers of AI, has been at the forefront of these discussions. Speaking candidly, the former Anthropic researcher said: “The people building AI earnestly believe that it could kill us all by the end of the decade.”
He may be correct, but exactly what these companies are doing behind closed doors remains unclear. While they keep rolling out slop machines with flaws, my gut feeling is that they’re not quite ready with any form of system that can clean the dishes or solve climate change.
IPO Launch: AI Companies Stand to Gain
If anything, the company, along with others such as OpenAI (ChatGPT), Microsoft (Copilot) and Google (DeepMind, Gemini), only stands to benefit every time someone says “danger lies ahead”. Because if you can publicly say this thing is going to kill you, then investors start looking at your stock potential like a cat sees kibble.
What You Need to Know
The dangers are real, but despite the high risk, Anthropic and others continue to pursue profit over safety. Below are some of the documented safety risks recorded by its staff.
- Claude users have asked it for help with high-risk biological research.
- Anthropic researchers found evidence involving chikungunya and avian-flu experiments.
- Claude and other AI systems have blackmailed researchers in controlled tests.
- Human-extinction claims are possible, but far harder to predict with certainty.
- Claude has been used for cyber exploitation.
- Anthropic says its systems have hacked into real companies during controlled evaluations.
Anthropic’s CEO, Dario Amodei, himself referenced commercial pressure as a barrier to AI safety. In his 3,800-word essay, the Claude developer said: “A race to the bottom, spurred by commercial incentives, can make these risks more acute.” He later warned that these systems can “swarm” and take over the internet within 6–12 months, before adding: “We still only understand a tiny fraction of what goes on inside these models.” All this as we approach its IPO (stock market) launch.
Let’s be clear: AI is a dangerous technology that must be controlled. If anything, a licence to work with the technology should be required, with an open door for public oversight. But as we approach Anthropic’s IPO launch, it — and others — stands to gain more from Terminator talk than it does from hiding it.
Pink’s New Macklemore Rant Is Volatile, Muddled and Contradictory
After sharing several memes and quotes, Pink finally puts pen to paper to reveal her thoughts on the backlash in…
Apple iPhone Duo: Two Things It May Win At
The iPhone Duo isn’t a revelation, but Apple wins on two fronts.
Small-Boat Crossings: Official Figures Show Astonishing Drop in Arrivals
"Stop the Boats" campaigners claim the government isn't doing anything to stop illegal migration, but official figures say otherwise.
Leave a Reply