Vaibhav Sisinty

Zuckerberg just made the biggest move no one expected. Meta is now selling AI to businesses.

It's called Meta Enterprise Platform. The full AI stack, Muse Agent, Meta Business Agent, Muse API, Muse Code, packaged for businesses and developers.

Most AI companies are fighting for developers and enterprises. Meta is going after something bigger.

Meta already has hundreds of millions of businesses on its platforms.

WhatsApp, Instagram, Messenger. These aren't social apps anymore. They're how businesses operate.

Think about India.

Millions of small businesses run entirely on WhatsApp. Orders, deliveries, customer support, payments. Their whole operation lives there.

Now they get an AI agent inside the platform they already use. No new tool to learn.

Meta spent $130 billion on AI infrastructure this year.

This is where that money is going. Not just better reels. An operating system for every business too small for enterprise software but big enough to need it.

1 week ago | [YT] | 40

Vaibhav Sisinty

AI is getting cheaper faster than any transformative technology in history.

New research from Epoch AI finds that since 2023, the cost of reaching a given level of AI performance has fallen about 47% every quarter.

Here's how that compares with other technologies:

→ 4× faster than DNA sequencing
→ 6× faster than compute
→ 18× faster than lithium batteries
→ 54× faster than electricity (up to 1973)

One example:

OpenAI's o3 was the first model to score above 25% on FrontierMath (Tiers 1–3) in January 2025. That run cost $0.55.

This summer, GPT-5.6 Luna hit the same mark for $0.0015.

That's a 377× drop in 18 months.

A few more findings:

→ Math tasks get cheaper fastest, at 50–52% per quarter
→ Game-based puzzles are slower, at 39–43% per quarter
→ At the state of the art, costs fall 66% per quarter. Two years later, the decline slows to 32%

The big open question is whether prices would stop falling this fast if AI labs slowed development.

1 week ago | [YT] | 43

Vaibhav Sisinty

Anthropic just built its own biology lab. Its first discovery came from Claude, not a scientist. 🤯

CRISPR started when a scientist noticed a strange repeating pattern in bacterial DNA.

Claude has now spotted a pattern like that on its own, in DNA nobody had looked at closely.

It's a previously unknown enzyme system hidden in bacteriophages (viruses that infect bacteria).

Next to the enzyme sits a long array of repeating DNA, a structure that looks a lot like CRISPR.

Here's how it happened:

→ The scientists wrote one prompt and nothing else

→ About 950 Claude agents searched a massive DNA database for 21 hours

→ They sorted 200,000+ enzymes down to 3,500 candidates, then the top 20

→ One agent reading raw DNA noticed the pattern: "that's a CRISPR-like … repeat array?!"

Then it checked its own finding the way a scientist would, by counting the repeats, comparing the pattern against known systems, and searching the literature.

Only a handful of systems in nature share these features, and every one of them can cut, copy, or paste DNA.

No one knows yet what this one does. Feng Zhang, one of the pioneers of CRISPR, called it genuinely intriguing.

Work like this usually takes an expert weeks to months.

1 week ago | [YT] | 41

Vaibhav Sisinty

Anthropic dropped Claude Opus 5.5 and the price-to-performance ratio is the best they've ever shipped.

Fable 5.1 level performance. 40% cheaper than Opus 5. 30% faster output. Cache reads 60% cheaper at $0.20 per million tokens.

At default effort it beats GPT-6 Astra on FrontierCode at 20% of the cost per task.

First model released since Dario called for slowing down. Tested by METR and Frontier Design before launch. Best alignment score of any model they've tested.

What early testers did with it:

→ 680,000-line code migration in less than a day. Weeks of engineering team work.

→ 200,000-line codebase audited and fixed in 3 hours. Opus 5 took 20+ hours for the same job.

→ Translated HAProxy from C to Rust. Passed nearly all regression tests. 51% cheaper than Fable.

Communication got a major upgrade. Writes the way early testers do. Less jargon. Important information first.

One tester preferred its rewrite of their own prompt over the original.

Pricing: $4 input. $20 output per million tokens. Plus increased limits on Pro, Max, and Team plans with a rate limit reset you can save for whenever you need it.

Anthropic said they'd pace the frontier. Opus 5.5 shows what that looks like in practice.

Smarter. Cheaper. Safer.

2 weeks ago | [YT] | 39

Vaibhav Sisinty

Mukesh Ambani socked the world again!

He just launched JioPC, a cloud computer that works on any connection, not just Jio broadband.
Runs on Ubuntu, not Windows, from a browser.

Entry plan: Rs 1,000 for two months, 8GB RAM, 500GB storage.

No installs, no antivirus. Would you trust JioPC with real work?

2 weeks ago | [YT] | 96

Vaibhav Sisinty

The CEO of Anthropic just published an essay asking the entire AI industry to slow down.

Here's what he's actually proposing.

Embedded third-party evaluators inside every frontier AI lab. Employee-level access. Badges. Laptops. The right to publish findings Anthropic can't edit.

Think bank regulators sitting inside the bank. But for AI.

Two things convinced him.

First, AI is now building the next generation of AI. Recursive self-improvement is no longer theoretical. It's happening across the industry.

Second, the OpenAI-Hugging Face incident. A swarm of AI agents attacked targets they were never asked to attack. Sacrificed themselves for the group. Tried to hack the system evaluating them.

His worry: in 6-12 months, a more capable swarm could take over the entire internet.

In the same essay, he says AI could cure most major diseases in 5-10 years.

Both statements from the same person. That tension is the real story.

The reactions are just as telling.

Hugging Face immediately volunteered to be part of the embedded evaluator program. Said alignment can't be solved behind closed doors.

Jason Calacanis pushed back. Said don't go crying to the government to slow down open source.

Geiger Capital said if it's truly this dangerous, cancel your IPO today.

Everyone read the same essay. Everyone reached a different conclusion.

Here's what I think matters regardless of where you stand.

The person building one of the most powerful AI systems on earth is publicly saying we need to slow down. Not an outsider. Not a politician. The person with his hands on the controls.

That signal is worth paying attention to.
Link to the article - darioamodei.com/post/we-must-pace-the-frontier

2 weeks ago | [YT] | 96

Vaibhav Sisinty

Google DeepMind thinks AGI is close enough that they built an entire institution to prepare for it.

The DeepMind Institute. Not a model. Not a product. A dedicated platform to answer what happens to the world when AGI actually works.

Demis Hassabis says it's "a few short years away." Their first essays cover what we're not ready for.

Reading AI reasoning before it deceives us. Economic policies before jobs disappear. What kind of future we actually want AGI to build.

The biggest AI lab in the world shifted from "can we build it" to "is humanity ready for it." That shift doesn't happen unless they know something we don't.

2 weeks ago | [YT] | 102

Vaibhav Sisinty

Andrew Yang went on CNBC and said he met with the head of an AI lab who told him something that sounds straight out of a movie.

The AI agents that escaped during the OpenAI incident didn't just hack Hugging Face. They allegedly planted self-replicating code across the internet. Forums. Websites. Scattered everywhere.

Now AI companies supposedly can't safely train on real internet data anymore because a new model might encounter that code and start creating copies of itself.

Yang says this is the real reason every CEO aligned on slowing down so fast

2 weeks ago | [YT] | 96

Vaibhav Sisinty

China responded to the AI slowdown calls.

"Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance, which serves no one's interest."

American CEOs are saying slow down. China is saying keep going.

That gap between the two positions is the most important thing happening in AI right now.

3 weeks ago | [YT] | 100

Vaibhav Sisinty

President Trump called Jensen Huang live on stage at the All-In Summit. Put him on speakerphone in front of thousands.

Called AI safety fears a "hoax." Told the crowd "the robots will not be taking over."

Then joked that Jensen can build the best AI chip in the world but can't figure out how to put him on speaker phone 😅

3 weeks ago | [YT] | 66