You can now create AI agents directly from your phone.
Replit lets you create them without writing a single line of code and completely for free.
Just duplicate the template and you're ready to go.
Links below
Two situations:
1 I wave a co-worker to my monitor to show them a file I have open. "Is this right"?
2 I sit down someone at a table. They have 2 hours to respond. This is an exam. The stakes are high. "Is this right?"
The humans collaborator knows 1 vs 2. The LLM doesn't know if it's 1 or 2, benchmarkmaxxing over time pushes it to assume 2.
@karpathy I cancelled cursor yesterday for this reason. Constantly thinking 5 steps ahead in the wrong direction. I’m back to manually copy + pasting from chatgpt/claude/gemini for now.
@paimon2cool I'm starting to do some of this too. I have a script that packages all of the files of my project (which isn't a giant repo and fits just fine) uses `files-to-prompt`, then I start new conversation, copy paste, and ask a question at the end, and manually select appropriate model.
I'm noticing that due to (I think?) a lot of benchmarkmaxxing on long horizon tasks, LLMs are becoming a little too agentic by default, a little beyond my average use case.
For example in coding, the models now tend to reason for a fairly long time, they have an inclination to start listing and grepping files all across the entire repo, they do repeated web searchers, they over-analyze and over-think little rare edge cases even in code that is knowingly incomplete and under active development, and often come back ~minutes later even for simple queries.
This might make sense for long-running tasks but it's less of a good fit for more "in the loop" iterated development that I still do a lot of, or if I'm just looking for a quick spot check before running a script, just in case I got some indexing wrong or made some dumb error. So I find myself quite often stopping the LLMs with variations of "Stop, you're way overthinking this. Look at only this single file. Do not use any tools. Do not over-engineer", etc.
Basically as the default starts to slowly creep into the "ultrathink" super agentic mode, I feel a need for the reverse, and more generally good ways to indicate or communicate intent / stakes, from "just have a quick look" all the way to "go off for 30 minutes, come back when absolutely certain".
Rakhi Beyond Ties
At Bennett University, siblinghood isn’t just defined by roots; it's found in shared notes, stolen snacks, saved seats, and 2 AM calls.
Happy Raksha Bandhan to the bonds that go beyond family ties.
#RakshaBandhan2025#BennettUniversity#CampusBonds
🇮🇳 भारत का जवाब अमेरिकी टैरिफ पर
1. टैरिफ असर का विश्लेषण और रणनीतिक प्रतिक्रिया
2. कृषि, डेयरी, MSME और युवा हितों की रक्षा
3. वैकल्पिक वैश्विक बाजारों की तलाश तेज
4.नई द्विपक्षीय वार्ता की जल्द योजना
5.नीति‑निर्धारण और राष्ट्रीय हित में संतुलन का परिचय
I am deeply honoured to take over as the Chief Secretary of Uttar Pradesh. My heartfelt gratitude to the Hon’ble Chief Minister Shri Yogi Adityanath ji for entrusting me with this responsibility. I am committed to serving the people of UP with dedication and commitment.
Identification of Feature by placement on screen not by icon: I interacted with approx two dozen users of TikTok, Youtube, Whatsapp. I wanted to know what triggers them to share a piece of content but I noticed that they couldn't identify the "share" icon. 5/n
I intend to use this thread as Filed notes .I will keep updating this thread over time. Feel free to contribute your observations to it. So grab a coffee and read on. 4/n
I am writing this thread to (1) Validate: is it just me and my sampling bias or others also observe the same patterns? (2) As we build products for the next billion users ( NBU) these observations might help us build a more representative persona for our users. 3/n
I have spent the last few months in Bikaner, A tier 2 city in Rajasthan. Here I got to observe our users from a distance while still being part of context. Some of what I noticed was contrary to the ideas we have in the industry about our users.2/n