hRiTiK

You will never know how strong you are until strong is the only option left for you.

💯 Yeah

I hope they will disqualify those legends otherwise it will be a joke 🤣🤣.

🤣🤣

My girlfriend called me at 2am crying. She had seen a photo on Instagram of me and another girl at a party. She sent me the photo. I looked at it and I'm like, what? Only my nose looks like the guy in the photo! I keep telling her, “We're not the same person,” but she is not ready to accept it. She then forwarded the photo to my friends asking them to confirm. Even they were confused. Bro that really does look like you. Now, at this point, the only hope I have is my last line of defense - a Cosine Similarity Test. I know you guys are thinking, what the hell is this Cosine Similarity. Cosine similarity is a mathematical way to measure how similar two things are by treating them as vectors in space. Think of it like measuring the angle between two arrows - the smaller the angle, the more similar they are. In math, cosine similarity works like this: cos(θ) = A·B / (|A| × |B|) Where: - A·B is the dot product of A and B. - |A| and |B| are the magnitudes. Understanding the Scale (-1 to 1): - cos(0°) = 1 : Perfectly identical - cos(45°) = 0.7 : Partially similar - cos(90°) = 0 : No similarity at all - cos(180°) = -1 : Complete opposites Now let me prove to my girlfriend that the guy in the photo is not me. Let's say my facial features are Vector A and the guy in the photo is Vector B: Vector A = [2, 4, 6, 8] Vector B = [1, 2, 3, 4] Step 1: Calculate Dot Product Multiply each corresponding element and add them all up: A·B = (2×1) + (4×2) + (6×3) + (8×4) A·B = 2 + 8 + 18 + 32 A·B = 60 Step 2: Calculate Magnitude Take the square root of the sum of squares of each element: A = [2, 4, 6, 8] |A| = √(2² + 4² + 6² + 8²) |A| = √(4+16+36+64) |A| = √120 B = [1, 2, 3, 4] |B| = √(1² + 2² + 3² + 4²) |B| = √(1+4+9+16) |B| = √30 |A| × |B| = √120 × √30 |A| × |B| = √3600 |A| × |B| = 60 Step 3: Apply the Formula cos(θ) = A·B / (|A| × |B|) cos(θ) = 60 / 60 cos(θ) = 1 Cosine of 1 means perfectly identical. Congratulations 🎉, you just learned Cosine Similarity. Bonus: Why does AI/ML care about cosine similarity? Recommendation Systems: Netflix uses it to find movies similar to what you have watched. Image Recognition: AI systems compare feature vectors extracted from images to identify faces or detect similarities between pictures. Document Classification: Text classification systems use it to categorize emails as spam or not spam by comparing document vectors.

he is 26, skipped the phd route entirely, and already leads a mechanistic interpretability team at google deepmind. he helped build anthropic's research team early on and has mentored dozens of top researchers, but he talks about it all with this incredible, grounded humility. sometimes you hear a conversation that quietly resets how you look at your own work. today that was @NeelNanda5 for me. just writing down my biggest takeaways so i don't forget them: > He became a DeepMind team lead mostly by accident. The old lead quit, and he just said yes despite feeling unready. We gatekeep ourselves way more than the industry does. The downside of trying is almost always zero. > Stop treating AI like a search engine. He treats it like a chaotic coworker. He dictates hours of messy audio notes and uses "anti-sycophant" prompts. You literally have to tell the AI, "My friend wrote this, please brutally roast it," just to bypass the polite filters and get honest feedback. > We cold email the wrong people. Reaching out to famous founders or senior leads is usually a waste of time. Email the junior dev who got hired six months ago. They actually have free time, remember what it feels like to be an outsider, and actively want to mentor. > If you just need a task done, let AI write the code. But if your goal is to learn the underlying architecture, outsourcing it to AI ruins your understanding. Use it to build fast, but treat it as a strict tutor when you need to actually learn. > Research taste cannot be rushed. You can learn syntax in months, but the intuition for what makes a project actually valuable takes years. You have to train your brain like a neural net by doing a bunch of low-stakes, messy projects until the deeper intuition finally clicks. > The hidden cost of brilliant ideas. You can write the most elegant solution in the world, but if it requires tearing apart a massive, pre-existing tech stack, an organization will reject it. The best engineers write code that is frictionless for the rest of the team to accept. the biggest realization is just how much agency we actually have. you can just do things.

.@ni5arga had found security issues with the CBSE dashboard for this year's boards. The website had - Hardcoded master password in the frontend JavaScript enabled login and OTP bypass. - OTP validation was done entirely on the client side, making the OTP step insecure. - Missing route guards allowed direct access to internal dashboard routes without proper authentication. - Password change API did not verify the old password before allowing a reset. - Systemic IDOR allowed users to modify client-side IDs and act as other examiner accounts. A malicious actor could take over examiner accounts, bypass OTP, access answer scripts, reset passwords, and tamper with marks at scale. being the responsible one, he reported it to the authorities, and the authorities took 0 actions (some of them are still up), he posted this blog today about the issue. His reward was him getting locked out of this twitter and the blog being removed from twitter by authorities really saddening

PSA: I would be free from August mid till Nov mid (Around 4-5 months) I am open for work during this short time and would love to work on any roles open, in a remote setting I am in my 7th sem soon, and would have a CGPA of around 8.9/10 I am currently interning at Cisco and working on developing router software (mostly C and some py). In the past, I have worked with top researchers of this country (India) and have several publications (2 A* Findings, 3 A* Workshops, One A* in review atm ~ high chance to convert into publication). I have also helped build critical models which have been deployed to lacs of people in production. Open for anything, remote + good work environment + only for those 4-5 months I can/will do good work, that's guranteed RT + Share to amplify thanks :)

Tried building an image using elemental prompting. Basically you model each component as an element instruct llm to use its properties to build up the image with a sample of your own image. Interesting output, need to research more. Will share the details and prompt later.

how to understand a codebase from the first principles. (example with a real production codebase)

@ThePitchTalks Why so jealous?

Bring this kid Vaibhav Suryavanshi to Perth. If he even reaches double digits against that pace attack, I’ll delete my account forever and retire from cricket analysis!

For the love of god DO NOT start CPR on a conscious patient with a pulse or patient with a pulse. CPR is for cardiac arrest, not for ‘looking weak, feeling dizzy, or the pulse feels feeble. A feeble pulse is still a pulse. Which means the heart is still contracting & generating at least some effective cardiac output. Weak circulation does not mean absent circulation. Chest compressions are meant to replace a heart that has stopped pumping not to ‘help’ one that’s already working. #MedX

How can an iit not provide decent living conditions... I'm literally so disappointed at how indifferent they are to our problems. - topmost floor with direct sunlight in this hot weather..rooms stay hot 24/7 - water coolers have scalding hot water - no wifi - very limited space

Why Kṛṣṇa devotees aren't afraid of anything?

A fake doctor is a real danger

@DocPriyamMD What about this ?

Med students, listen up! Today let me explain how to analyze a Complete Blood Count (CBC) report step by step. A CBC is more than just numbers; it’s a clinical narrative. Let’s break down this real patient’s report together. 👇

@ayaz_nagouri @cneuralnetwork Kabhi face pe kisi ko jhaatu bola h kya ? Online toh sab bhokte h par tere jaise bacche real mein moot dete h.

@23_dec_2005 @cneuralnetwork Jhatu 😂

tomorrow rcb vs kkr going to blr as well, will wear the russell jersey 😁

I don't know from which logic he threw the ball on keeper's end when bhuvi was already ahead.

RCB vs MI match is completely fixed. This mdc franchise needs to be banned from IPL forever

@ParthSRajawat 😂😂

@23_dec_2005 Bade dino baad prji

I was also shocked watching this and it was intentional, because the ball was not in that much speed as well and as a professional cricketer you are not able to collect it is a SUS.