Ashish

Have u planned and architected your own home ? any good read to begin with. an agent would be a plus.

In Parliament today I proposed optional Joint filing of Income Tax Returns for married couples. Family A Both spouses earn ₹10 lakh each. Total household income is ₹20 lakh. Tax: zero Family B One spouse earns ₹20 lakh. The other stays home to raise their child. Total household income is ₹20 lakh. Tax: ₹1.92 lakh The only difference is how the salary is split between the two spouses. One roof. One kitchen. One household budget. But when tax time comes, the family disappears. The tax system sees two individuals. A husband and wife become strangers. No clubbing of income or rebates. In Parliament today I proposed optional joint filing of Income Tax Returns for married couples, so families with uneven incomes are not unfairly penalised. If implemented, then Family A and Family B both will pay Zero Tax.

RDF vs Property Map for Enterprise Knowledge to Power SRE Agents ? What are you using ? #sre #ai #rdf #graph #EnterpriseAI

@DhravyaShah Thanks for the insight, I had similar suspicions but good to see the direction industry is heading to. this further proves @supermemory is in right direction

been building in this space for years now, and have followed nishkarsh for years as well - congrats on the launch! since this is in the same space we're building in, i dived deep into it and have thoughts. the launch itself is very hype-y, and is meant to trigger rage bait 1. it's positioned as a database, but is almost a @supermemory-like system 2. their example of "vector dbs" not being able to do this, is really a question of "embedding models". and embedding models have superpositions, they are cheap and are easily able to infer differences between them. it's not hard to ask claude to do a mini experiment to prove this (attached below). What does matter is: is it able to track how knowledge evolves? time passes? this made me curious so i read their paper 3. their research paper is hardcoding and gaming the benchmark by different prompt for every category!!! (see image below). If their benchmarking is fixed, supermemory will remain the SOTA. 4. they reinvented contextual retrieval paper by Anthropic from 2024 and called it "the orphaned pronoun paradox" 5. they mention they use a custom "in-memory vector store" = at about 500GB, you will have to pay more than $10k for just the RAM. 6. inference is run too many times in the pipeline - which means for every LLM token you ingest, you will end up paying 5x more than token cost for the graph + contextualization + storage. 7. latency and cost numbers were never reported. My hunch is because of the architecture, the latency will struggle at scale. but i can't tell - their product is behind demo gate. 8. the benchmarking code is not OSS (from what i can tell). not replicable + who knows how much context they are injecting into the model? what's the K? 9. inorganic, undisclosed ads (just read the quote tweets). influencer accounts with 400k+ followers all saying the same thing. people keep getting away with this @nikitabier lol i'm all in for healthy competition and progress in this fields, enjoy seeing good work being done by others. but its easy to just say things. "no one will check." playing the game the right way is hard, and everyone's just saying whatever they can to impress people. TLDR is: you should use this if you want to spend 2-5x more for no real marginal improvement and enjoy unhealthy research and business practices. attached: 1. experiment to disprove hypothesis of vector dbs not understanding grey vs grey 2. one of their prompts, which just says "say i dont know". they scored 100% :)

@contextkingceo Interesting , this week I am starting a project that builds RAG over million of records. looks like I should rethink my strategy! any good read to start with ?

We've raised $6.5M to kill vector databases. Every system today retrieves context the same way: vector search that stores everything as flat embeddings and returns whatever "feels" closest. Similar, sure. Relevant? Almost never. Embeddings can’t tell a Q3 renewal clause from a Q1 termination notice if the language is close enough. A friend of mine asked his AI about a contract last week, and it returned a detailed, perfectly crafted answer pulled from a completely different client’s file. Once you’re dealing with 10M+ documents, these mix-ups happen all the time. VectorDB accuracy goes to shit. We built @hydra_db for exactly this. HydraDB builds an ontology-first context graph over your data, maps relationships between entities, understands the 'why' behind documents, and tracks how information evolves over time. So when you ask about 'Apple,' it knows you mean the company you're serving as a customer. Not the fruit. Even when a vector DB's similarity score says 0.94. More below ⬇️

THIS:

@naval i read the same on @paulg blog “How to do great work”!

Great work is fractal in nature - elegant from a distance, exquisite in the details.

Excited to introduce PayPal World 🌎, a first-of-its-kind global platform connecting the world’s largest payment systems and digital wallets. Our team has been working with partner wallets around the world to change the game for billions of consumers and merchants everywhere. Let's break it down:

@springrod @slimslenders @Docker @kotlin @springcentral Established distribution (docker) catching up with the innovation (hugging face)!

Docker has a lot to offer for the agent era, with Compose, local model support and MCP gateway. Happy to be working together with them at Embabel. @slimslenders @Docker #Docker #DevTools #AgenticApps #Java @kotlin @springcentral #AgenticAI docker.com/blog/build-ai-…

30 feels like something. I can officially say it has some weight. #turning30

AI PROMPTING → AI VERIFYING AI prompting scales, because prompting is just typing. But AI verifying doesn’t scale, because verifying AI output involves much more than just typing. Sometimes you can verify by eye, which is why AI is great for frontend, images, and video. But for anything subtle, you need to read the code or text deeply — and that means knowing the topic well enough to correct the AI. Researchers are well aware of this, which is why there’s so much work on evals and hallucination. However, the concept of verification as the bottleneck for AI users is under-discussed. Yes, you can try formal verification, or critic models where one AI checks another, or other techniques. But to even be aware of the issue as a first class problem is half the battle. For users: AI verifying is as important as AI prompting.

Sam Altman says the perfect AI is “a very tiny model with superhuman reasoning, 1 trillion tokens of context, and access to every tool you can imagine.” It doesn't need to contain the knowledge - just the ability to think, search, simulate, and solve anything.

i must study war so that my sons can study mathematics, economics & technology, and so their sons can study poetry, art & philosophy.

Can u we have local currency support for "TOTAL TOKENS" in langsmith @hwchase17 @langchain I want to see in usages in INR

@paraschopra @lossfunk @HiSohan What’s the intent of this ? How will this progress tech? Or tech in India ?

Announcing India 🇮🇳 ML tracker, an open source @lossfunk project built by @HiSohan to track papers accepted at top AI/ML conferences by Indian authors.

@svembu Sir wait a little, let this road or even the pradhanmantri road come to us first 😁

For the first time ever, I took the Chennai-Vikravandi-Kumbakonam-Thanjavur highway that runs through my native village Chidambaranathapuram and the exit sign makes my small village feel "important"! Now I want to see an electric train station too!

All of the billions spent on AI innovation just so we can get these beautiful cat videos

So this morning I received a msg from paypal that they blocked unusual activity on one of my credit cards. I check on my bank account and sure enough there is a small amount this morning from someone in India. No idea how Paypal did it, but I'm glad they did.

(1/2) Hi My name is Rohan Mehta and I will try to make this as easy to understand as possible. I have been employed with Ministry of External Affairs for more than 8 years now. During this period, I have served in Embassy of India, Moscow and Consulate General of India, Sao Paulo. While working in Sao Paulo, I learned that the Consul General (head of Consulate) was a totally corrupt lady Manisha Swami and she used to harass all employees. Her modus operandi was simple: she used to come to office for some time and abuse and shout at the employees and then leave for the whole day. She was also misusing official resources. There was another Consulate employee (Adarsh Mishra) who was harassing female employees and not working at all. Before I move further, let me tell you the salaries of both of these employees which is effectively taxpayers' money(USD $7000 and USD $5200 per month approx. respectively). As Consul General learned that I was the only one who could complain, she nudged another employee to make a fake sexual harassment complaint against me. The goal was that I either back off or probably commit suicide as a result of the fear they were trying to create in my mind. But why would I have fear when I had done nothing. The name of the lady who complained is Gayatri Gosain. This lady used to tell me that she used to treat me like an elder sister. And I gave her all due respect like you would to your sister. By complaining against me, she did an extremely cheap thing. But when this complaint reached Ministry of External Affairs and when she was asked if she wants to proceed with the complaint she backed off (either because of shame because she knew that she made a fake complaint or because she knew that Complaint Committee takes action against the complainee when it is fake). Now also, I have all documents to prove it was fake. I will post an image which shows how much harassment I was facing while I was in Sao Paulo. Also, Adarsh had also harassed Gayatri Gosain and the first 3 minutes of one hour recording on my Facebook post(link is available below in the tweet about sequence of posts) have proof of the same. Now let's move further. I eventually made a corruption complaint on 5 January 2024. I addressed it to Indian Ambassador in Brazil Suresh Reddy and Administration head in India B Vanlalvawna. I informed about the corruption and also about the sexual harassment instances. But instead of taking any action against the culprits, I was recalled to India and was asked to travel back on 9 January 2024. I intimated that I was not vaccinated against yellow fever so I maybe allowed to take vaccine and travel after 10 days once the vaccine becomes effective. Regardless, I was given an exemption certificate to travel without vaccine(this was a clear safety hazard). I am attaching a copy of the certificate here. The guy who signed this certificate was also harassing another Brazilian lady and a recording of the same is available on my Facebook post. dropbox.com/preview/Vaccin… I returned to India and asked about action taken on my complaint. I was told that no action will be taken against it. In a short while, Ministry opened a fake case against me alleging that my fiancée(who is Brazilian) was living with me in Brazil. My fiancée was not living with me in Brazil and I intimated the same. I also intimated that I was facing harassment and I even submitted a copy of corruption complaint to Vigilance Section. They called me two months after my reply and tried to coerce me to change my statement and write that my fiancée was living with me. This coercion was openly illegal and anyone who wants to know what happened can listen to the following recording. dropbox.com/scl/fi/qp9z99d… This recording is 40 minutes long and in the second half the Undersecretary from Vigilance Section is repetitively trying to force me to change my statement. In the first half, another personnel from Vigilance tells me that generally action is taken against people who complain about corruption. @ShashiTharoor @Ram_Guha @Wangchuk66 @SupriyaShrinate @Pawankhera @mohakmangal @foodpharmer2 @thevirdas @theskindoctor13 @MANJULtoons @_YogendraYadav @thewire_in @SauravDassss @sandeep_PT @rakhitripathi @jawharsircar @ajitanjum @barkhatrehan16 @satishacharya @malpani @AtishiAAP @SanjayAzadSln @raghav_chadha @MahuaMoitra @TheDeshBhakt @bhagatram2020 @prakashraaj @apnarajeevnigam @ravishndtv @mynameswatik @Sydusm @mediacrooks @newslaundry @MnshaP @AbhinandanSekhr @BBCIndia @AAPforNewIndia @Amockx2022 @sanket @rantinggola @ReallySwara @mohitlaws @suchetadalal @nitishrajput @Harvard @Stanford @Yale @Columbia @mehdirhasan