Google has published a paper that might end the transformer era.
For the last 7 years, every major AI, ChatGPT, Claude, Gemini, has been built on the exact same architecture: The Transformer.
But Transformers have a fatal flaw.
To remember context, they have to process every single word against every other word. It’s called quadratic complexity. As your prompt gets longer, the compute cost explodes.
The alternative is the old-school RNN (Recurrent Neural Network). RNNs are incredibly cheap and fast, but they have a fixed memory size. If you give them a long document, they get amnesia.
Until today.
Google researchers published Memory Caching: RNNs with Growing Memory.
And it fixes the biggest bottleneck in AI.
Instead of an RNN having a fixed, rigid memory that constantly overwrites itself, Google gave it a "save" button.
The technique allows the RNN to cache checkpoints of its hidden states as it reads.
The memory capacity of the RNN can now dynamically grow as the sequence gets longer.
They built four different variants, including sparse selective mechanisms where the AI actively chooses exactly which checkpoints matter most.
The results rewrite the rules of efficiency.
On long-context understanding and recall-intensive tasks, these new Memory-Cached RNNs closed the gap with Transformers.
They achieved competitive accuracy without the explosive, quadratic compute cost. It perfectly bridges the gap between the cheap efficiency of an RNN and the massive capability of a Transformer.
We have spent billions scaling Transformers because we thought they were the only way an AI could remember a long conversation.
But Google just proved we don't need to process the whole history every single time.
We just needed a smarter cache.
@AndrewCurran_ How does this help an average citizen ? Can we apply for some scholarships to learn these AI courses or certifications free of cost and where do we apply them ?
This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨
Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser...
The numbers are hard to believe:
> $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built
> Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet
> Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit
For context: these are the kinds of findings that previously required elite security researchers working for weeks.
Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning.
The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing:
A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.
LiteLLM HAS BEEN COMPROMISED, DO NOT UPDATE. We just discovered that LiteLLM pypi release 1.82.8. It has been compromised, it contains litellm_init.pth with base64 encoded instructions to send all the credentials it can find to remote server + self-replicate. link below
🦔 Jason Calacanis says his company hit $300/day per agent using Claude's API at only 10-20% capacity, which scales to around $100,000/year per agent. Chamath Palihapitiya added that he's now asking "what's the token budget for our best devs?" and said AI-assisted developers need to be at least 2x as productive just to justify the cost. He said this is actively happening inside his company or he'll run out of money.
My Take
This was always the obvious trajectory. AI providers subsidized usage to drive adoption, and now the subsidies are ending. The consumer plans are likely loss leaders subsidized by VC money, and the gap between what individuals pay and what it actually costs to run these models is closing fast.
I'm struck by the surprise from people who should know better. These are sophisticated tech investors just now realizing that running agents 24/7 burns through tokens at rates that dwarf human salaries. A human engineer runs on coffee, remembers context from years ago, builds institutional knowledge, and doesn't rack up exponential costs the longer they think about a problem.
Agents waste tokens constantly, researching and validating things that don't need validation, spinning up subagents for simple tasks when a straightforward approach would work fine. The companies that fired engineers to replace them with AI agents are learning that you can't negotiate with an API bill the way you can renegotiate a salary. And unlike employees who might stick around during a rough patch, the meter just keeps running.
Hedgie🤗
In 1999, Buffett explained why most people never get rich.
He proves you are walking around with a $500,000 asset right now.
Save this rare footage of this before you lose it.
Google loves keeping valuable websites hidden from you.
Here are 10 incredibly useful AI websites you probably didn’t know existed.
Trust me, you’ll want to bookmark these 🔖