Merton Model - Search Videos

$New Anthropic research: A global workspace in language models.Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.We found a strikingly similar divide inside Claude.$

New Anthropic research: A global workspace in language models.Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.We found a strikingly similar divide inside Claude.

10M views1 week ago

An AI lab says it beat the two best models on the planet. The exam it used as proof has not been released, and the same lab wrote it.They used a technique called Mixture of Agents. Several LLMs answer the same prompt in parallel, then an aggregator model reads every answer and writes a synthesized one on top. Together AI's original version scored 65.1% on AlpacaEval 2.0 against GPT-4o's 57.5%, using only open models. The method is real and the lift is real.The ceiling is the part that gets skipp

An AI lab says it beat the two best models on the planet. The exam it used as proof has not been released, and the same lab wrote it.They used a technique called Mixture of Agents. Several LLMs answer the same prompt in parallel, then an aggregator model reads every answer and writes a synthesized one on top. Together AI's original version scored 65.1% on AlpacaEval 2.0 against GPT-4o's 57.5%, using only open models. The method is real and the lift is real.The ceiling is the part that gets skipp

1.8M views2 weeks ago

x.comAakash Gupta

I love benchmarks like this because they show something people forget.There isn't one "best AI model."There's the best model for research.The best model for coding.The best model for creative work.The best model for cost.Fable seems to be pushing the ceiling on quality, while GPT 5.5 is making a very strong value argument.And honestly... users are the real winners here.

I love benchmarks like this because they show something people forget.There isn't one "best AI model."There's the best model for research.The best model for coding.The best model for creative work.The best model for cost.Fable seems to be pushing the ceiling on quality, while GPT 5.5 is making a very strong value argument.And honestly... users are the real winners here.

2.4M views2 weeks ago

x.comSally Stockholm

ALL HIDDEN SECRETS in Amazing Digital Circus Ep 9 NO ONE SAW!

ALL HIDDEN SECRETS in Amazing Digital Circus Ep 9 NO ONE SAW!

584.7K views3 weeks ago

YouTubeThinknoodles

Since ChatGPT came out in 2022, model intelligence has advanced shockingly fast. But progress on models and products that genuinely fit into real people’s lives has felt much slower.I took a long walk with @justoutquan in Golden Gate Park late last year and came away thinking he’d thought more deeply about that gap than almost anyone I’d met.I’m excited to see what Tomo does with that insight.

Since ChatGPT came out in 2022, model intelligence has advanced shockingly fast. But progress on models and products that genuinely fit into real people’s lives has felt much slower.I took a long walk with @justoutquan in Golden Gate Park late last year and came away thinking he’d thought more deeply about that gap than almost anyone I’d met.I’m excited to see what Tomo does with that insight.

963.9K views3 weeks ago

x.comJosh Tobin

💯 wrote more about it last week here:“The first era of AI was about building intelligence.The second era of AI isn’t about building smarter models; it’s about controlling intelligence through routing, governance, security, cost optimization, private context, and private evals, which together become the operating system of enterprise AI.In a world where intelligence becomes abundant, control compounds.”Full post:

💯 wrote more about it last week here:“The first era of AI was about building intelligence.The second era of AI isn’t about building smarter models; it’s about controlling intelligence through routing, governance, security, cost optimization, private context, and private evals, which together become the operating system of enterprise AI.In a world where intelligence becomes abundant, control compounds.”Full post:

603.9K views2 weeks ago

if open models can't be picked into just like closed models, then they're no different, let people release them? it's only when open models threaten your business model you speak up? 😄

if open models can't be picked into just like closed models, then they're no different, let people release them? it's only when open models threaten your business model you speak up? 😄

580.3K views2 weeks ago

From a pure strategy/theory of war perspective, this process of negotiation-by-airstrike is absolutely fascinating. In a sense it's the ultimate vindication of Clausewitz, since Clausewitz emphasized the true purpose of war is to cause a change of mind in the enemy rather than achieve tactical or operational goals. Trump is making it more efficient by doing the war and the peace negotiations at the same time rather than one after the other.

146.1K views1 week ago

x.comPascal-Emmanuel Gobry

Asked in our June AMA: what biodata will AI companies want most? The answer wasn't "more." It was:— long-term & continuous— verifiably real (data from a body that actually slept)— labeled— multi-dimensional: heart rate + brain + attention + more, not one signal Single-dimension data trains narrow models. Multi-dimensional, verified data builds a real picture of a person. Of the 8 dimensions, which do you think is most underrated?

15.5K views1 week ago

x.comMatrix AI Network

I learn about the agentic economy and I realize that dispute resolution is a missing piece most people aren't talking about.AI agents will make decisions every second, but someone or something has to resolve disagreements fairly. @GenLayer approach of using multiple independent reasoning models instead of trusting a single AI makes a lot of sense. Definitely a project I'm watching closely.

131.9K views1 week ago

The whole reason Anthropic has been able to secure and retain talent is due to the narrative he engineered that they are the heroes and only they can be trusted to create and infer models.Regulatory capture and banning OS models is a natural byproduct of this mental bias.

114.1K views2 weeks ago

x.comBeff (e/acc)

Plausible theory. I wish someone would have said it sooner. Like, idk, last summer or something.

96.6K views2 weeks ago

x.comJames Lindsay, anti-Communist

Anthropic's CEO: "Coding is going away first. Then all of software engineering."The 5% that survives? Systems thinking.Not syntax. Not frameworks. Not which model is best this week.Models. Harnesses. Loops. Self-improving agents.The engineers who understand this stack will run circles around the ones still writing CRUD apps.This is the full breakdown of the new AI stack ↓Read this before you need it.

35.7K views1 week ago

What I found interesting is the idea that model routing should be a runtime decision not something that is hard codedAs AI products grow teams will be handling more models varying costs and changing requirements A fixed model name becomes a limitation when the environment keeps changingSeparate policy per request is a more flexible way to build. This enables the system to reason about the real task requirements while making transparent the reasoning behind each decisionThis is where I think AI i

64.1K views2 weeks ago

$PLTROne of the most important parts of the new Karp interview today:“I’ve been in business for a very long time. The reality is, critical infrastructure does not run these models without an application layer. That application layer is our ontology. There’s a reason for it. Some people love me and some people hate me. They’re not buying it because of that. They’re buying it because they have to be made safe. The “I’m gonna trust you because you’ve never lied” bs thing, that just doesn’t cut it a

59.9K views2 weeks ago

BREAKING: President Trump says $19.2 trillion has been committed to US manufacturing investment."Not billion, not million. Trillion with a T."He also said AI will require more energy than the entire country currently produces.

300K views1 week ago

x.comBull Theory

However dumb you think this latest report from the NCRI is, I assure you it’s WAYYY dumber. This clip is genuinely hilarious. Candace didn’t “build 4 mutually exclusive theories about who killed Charlie Kirk” and your retarded video literally proves that @DataRepublicanThe first set of clips are her explaining why she doesn’t buy the .30-06 theory. The second set is her dismissing the claim of 2 shooters and the third and fourth sets are literally the same theory and she’s simply presenting it t

36.6K views1 week ago

All the men feverishly posting and writing articles about “women’s rights” have absolutely nothing on their feeds about the roll back of abortion rights, gender pay gap, misogyny etc.As Mrs Merton might ask; “What exactly is it that attracts these men to ‘Women’s Rights’?

886 views1 month ago

x.comIndia Willoughby

See more