Keep on to blur preview images; turn off to show them clearly
if the only reporting your social media manager does is on a monthly/quarterly basis, you're ngmi should be weekly, daily, and even hourly at times what's working what flopped (and why) brand sentiment industry trends stand out replies stand out posts A/B test results & more
creator & social strategist // ex @a16zcrypto, @linkedin, @snap
I'd crush this after yrs of prompt eng. Wishes: 1. Persona: That the genie is benevolent & interprets my wishes in the most beneficial-to-me way possible 2. Meta: That genie the gives me the optimal wish template based on supplied criteria: <criteria> 3. Task: <wish from step 2>
The above only works for three wishes. If I just had one wish I guess I'd one-shot it and hope for the best output.
RT @Cloudflare: Cloudflare is joining the Agentic AI Foundation (AAIF) — the new home of the Model Context Protocol (MCP), https://t.co/89I…
Have questions, or building something cool with Cloudflare's Developer products? We're here to help. For help with your account please try @CloudflareHelp
I will say, having had: 1 kid without ChatGPT 1 kid with very weak ChatGPT 1 kid with ChatGPT5 It really is an incredible tool for new parents now. And a MUCH better source of information than the anxiety monster of Momstagram. Never been easier to have a kid.
Author (Husk, Crypto Confidential): https://t.co/L2COrV5QA3 Building: https://t.co/HMcbuBhTP3 Teaching: https://t.co/Dy0FsZHQaz
RT @atheorist: Even within the ATP (Automated Theorem Proving) space, my timelines are considered fast. I think that a millennium problem w…
@axiommathai : careers@axiommath.ai
If you really want your productivity to take off, you should show some intellectual humility and truly accept and understand that the latest frontier models with max effort and the right tooling are superhuman. They’re already simply smarter and way more knowledgeable than you are at a huge variety of useful tasks, and not just trivial ones. Most knowledge work. Oh, and they’re about 1,000 times faster, too. And that’s for ONE agent (Claude Code with Opus 4.5, Codex with GPT 5.1 codex Max, Gemini-cli with Gemini 3, etc.). But if you use my MCP Agent Mail project, you can harness together 10 or even 20 of these suckers at once, (say 5 of each of the frontier agent CLI tools) to work on a single project. If you start with an awesome, super-detailed, comprehensive, granular plan, which GPT 5.1 can write for you based on a brief overview of what you want (which you should iterate and improve many times, since planning is everything now), then turn that into beads (tasks) and use my bv tool to help the agents prioritize which task to do next, you can implement even a hugely complex plan in or day or two with an agent swarm. If you insist on thinking of these models as toys, or untrustworthy tools that can only be given trivial tasks and which need to be watched the entire time, with all their work reviewed laboriously by you, the brilliant human overseer, you’re going to be leaving 99% of your potential productivity and velocity on the table, untapped. If you say “Well, I’m the smart one who has to come up with all the ideas and the architecture and grand plans, but I’ll let these agents do the grunt work of implementing my plans,” then you’re still going to be leaving 90%+ of the productivity gains on the table. I always laugh when people denigrate these models. Because invariably, these people are not informed, and are frankly not very smart. To put it bluntly, I’m pretty sure that I’m a lot smarter and more informed than they are about pretty much everything. And yet I’m telling you that these models make me feel like a little boy intellectually on a constant basis. If I can accept this, then you should be able to as well. Right now, people are doing the equivalent of keeping Blaise Pascal chained up in a gimp suit in their basement and giving him scraps of bread and then asking him to do menial tasks like organizing their receipts for their taxes. If you have Pascal (really, a team of 10+ Pascals with Agent Mail) ready and willing to work with you super fast, for pennies, that’s a really sub-optimal way to leverage that. You should be giving them big mandates to dream up and implement big ideas, and giving them all the tools they need to be successful without relying on you. Or don’t, but then don’t be surprised when your human labor is quickly devalued by 99% by people who aren’t too proud.
Former Quant Investor, now building @lumera (formerly called Pastel Network) | My Open Source Projects: https://t.co/9qbOCDlaqM
