Sagar

Thank you govt for the incentive for hard workers

This is what free capital markets looks like the consumer always benefits at the end

@zephyr_z9 No one cares

Gemini 3.5 Pro is delayed

@SumitM_X there is no comparison

BKC or Bellandur ?

In this AI age life come faster

Boris here is my thesis and need your experienced POV to judge this? The AI stack in any org is a 3 layered cake: 1. Context 2. Tools 3. Orchestrators Context is owned managed and built over time using vector dbs, ASTs etc. This is where the magic lives. Orgs must spend their time in this layer. Tools are the Clouds / CICD pipelines / DBs etc. this also doesn’t change much but has high impact on the overall health of the organization’s output. Org must build this once and build well. Orchestrators- it’s changing everyday. There is no point fighting for the right approach to ship software or artefacts. Harness / Loops / Skills etc lives here. Org members should have full autonomy to choose what they do here in this layer. I’m an AI consultant and the above framework is working well for my clients. Do you see any areas of improvement/ how would you think about it in enterprise context?

I talk to engineers at other companies every day and hear the same thing: one person is 10x'ing their output with Claude but the rest of the org hasn't caught up. Watching teams adopt AI, I keep seeing the same 4 steps. I mapped them out here: Steps of AI Adoption claude.ai/code/artifact/…

@cerebras This is awesome. Thanks authors

All that drama for Fable just to get beaten up by Kimi and Sol

@ats1999_er Like we remember Steve Balmer and Tim Cook

@code_sagar Wrong, everyone would remember him. Google CEO is big thing in itself.

No shareholder is going to remember him

the next set of events clarifies why you shouldn’t piss off the Gods

She didn’t see me

@airtel is a black spot in my life. The problem is I have taken almost every service that they have to offer in there Black plan Now switching cost has become high. Every morning in my most productive hours the Internet will be gone But it will show some BS names like Airtel Fastlane etc. Like wtf man. So guys be careful and don’t subscribe to one provider fully like me. Distribute it.

🏴󠁧󠁢󠁥󠁮󠁧󠁿 Inglaterra: 4-2-3-1 en rombo, Kane de falso 9, Bellingham llegando desde segunda línea, Pontas invertidos, Presión alta, Rotaciones en el mediocampo 🇦🇷 Argentina: El que no se la pase a Messi es gay

The referees were biased FIFA is biased Trump asked FIFA to make ARG win +100 other BS But remember kids this man named Lionel Messi is the difference

@cneuralnetwork @airtelindia @airtelindia is a black spot of my life as well. The home Wi-Fi is as shitty as it could get

since last night my mobile shows 5g+ but literally 0 internet is working, it shows 5g but no internet i have raised complaint many times in airtel app but they are not fixing anything. how do I fix this issue? @airtelindia please help me I can't even go out because of this!!!

You don’t know what is a handball do you?

Product Definer and Product Builder are the only two roles that will remain don’t optimise for DSA, Java and what not. Save this for later.

@ashwingop True

If McKinsey or BCG consultants, hired for their intelligence, tries to do what frontier models are doing right now to enterprises, they’d be sued out of existence. Yet, by "renting intelligence" from model providers without ensuring that they exclusively own their context, leaders are actively funding their future competitors. When an enterprise hires consultants like Mckinsey, they have to sign strict, enforceable, non-competes and exclusivity clauses. The consultants can't just learn from their clients and give it to their competitors or start efforts internally to compete with their clients. With the use of frontier models in enterprise there is no such understanding. Enter the "Reverse Information Paradox" that @satyanadella talks about. These early enterprise adopters are going to pay for AI with money, and then they pay a second time their proprietary knowledge that is arguably far more valuable. We already have two examples Cursor and Figma with Anthropic. What makes matters worse is that the more one wants from the AI, the more internal context you have to feed it. Every prompt, every tool an agent uses and much more importantly every correction a user makes when the AI inevitably messes up is "intelligence exhaust" the model provider will happily absorb. The cruel irony in all this is that the enterprises are "renting intelligence" while actively giving the model providers all their institutional know-hows in the form of context they should be owning. If you wouldn't allow the consultant who you hired to go work for your competitors or worse yet become your competitor themselves, why are you blindly doing the same thing with model providers? The business geniuses have to ask themselves or are they so AGI pilled that they forgot fiduciary common sense. The solution isn't just exclusive models or open-weights, though they would definitely help. The long-term, robust, solution involves redesign of the AI stack by separating "intelligence" from "context". Specifically, enterprises have to treat models, foundational or open-weight, as they do consultants. Rent raw cognitive horsepower by the token, decouple it from the orchestration layer, the private evals, and most importantly own their organizational memory/context. The legal firewalls from the consulting world need to be adapted into architectural firewalls by decoupling model, orchestration and memory. Rent intelligence. Own context.

When the boss says protect your context you should do it