My 6839 Bookmarks & Likes

A curated list of liked, bookmarked, and retweeted posts from Twitter.


I created my LinkedIn just 16 weeks ago. It’s still the absolute easiest algorithm to crack by a longshot. In a relatively short amount of time, I’ve recorded over 680,000 impressions, made more than 4,200 connections and added $14,000/month in revenue. Here’s exactly how I did https://t.co/JKzAFTqRGj

RIP cheaters. Our tool catches interview coder's invisible screen! https://t.co/4tczagm13b Demo below 👇 https://t.co/JSXBBRMjdH

In 2023, a humble blogger discovered discrepancies in photographs taken by the Dana Farber Cancer Institute that led him to the realization that they were fabricating data. Yes, a random dude dethroned one of Boston's biggest research sources by following a trail of breadcrumbs. https://t.co/8gawTwRSch

Introducing SpeakerKit State-of-the-art on-device speaker diarization: - 10 minutes of audio processed in 3 seconds - 10 megabytes in total - 6-year-old devices supported Details and links to the demo app are in the thread. https://t.co/BhhscrOyIi

the more i teach the more i am convinced that the only thing that separates the "smart" students from everyone else is whether or not they think the material is cool

This guy literally shares the best AI app-building advice in one page https://t.co/4WkyagSIz2

to me, the greatest sadness of all is the banality of most people's lives i'll spend time with my non-elite friends, and it's clear that their lives are devoid of heroism and lore they lock in/clock in, but they fight no great battles, they dash no great foes they drop by a…

I did it. I built "Cursor for Writing". Introducing Lucid ✨ - Write at the speed of thought - Pull in context from anywhere - Chat + Deep Research https://t.co/9Ftu2UkkJE

Introducing our new family of state-of-the-art AI models: Phoenix-3, Raven-0, and Sparrow-0. Together they bring Conversational Video Interfaces (CVI) to the next level, and power Charlie, our new demo persona 👋 https://t.co/vaqX9InvwX

Introducing the world's best OCR model! https://t.co/Mi04wgj6cM

Cursor rules are the secret to making AI actually understand your codebase. Let the AI create its own rules by extracting patterns from your docs and existing code. Then watch as it follows your project's conventions perfectly. No more explaining your standards over and over https://t.co/Te12EzX0Rm

Introducing Manus: the first general AI agent. Try Manus today and see the future of human-machine collaboration: https://t.co/IKBPiiHsta https://t.co/z3C8fUK3eN

SaaS marketing toolkit: • Postbridge → schedule content • Meta ads → run paid ads • CapCut → edit video • Outrank → seo on autopilot • Instantly → send cold emails • Enrichlead → get leads from traffic • ChatGPT → everything else

Once, A phd guy (from the US), in my connections, had reviewed my resume and gave me very detailed feedback on it. I applied the feedbacks and reached 94 on ATS and got some really good responses from the recruiters. Full conversation in the comments. 👇 https://t.co/Al0Ut9pU9O

"AI Agents are cron jobs", but they can also be complex step-function workflows. You need durability, retries, steps, realtime feedback… We announced @vercel Fluid compute which makes it cost-efficient to stream long responses that wait for LLMs. @upstash offers a great https://t.co/r0V6pCabjb

We made an agent that can generate @3blue1brown style videos about *anything*. We've also released a benchmark for animated educational videos so we can systematically improve this tech! Some notes below: https://t.co/zVf1eHRNra

Introducing NextGenAI—a first-of-its-kind consortium uniting 15 leading institutions using AI to advance research and education, founded with $50M in funding and tools from OpenAI. https://t.co/6rybRcJvGG

Announcing the LM Studio SDK for Python and TypeScript! 🔨 As a part of the SDK, we also introduce our first agent-oriented API: the `.act()` call. You provide a task and tools, and the model goes on its own autonomously for multiple turns until it accomplishes the task. https://t.co/mkq9I6LsMB

Accelerate data science workflows with Google Colab's new Data Science Agent. This agent uses Gemini to act as your coding partner: upload your data, define your analysis goals, and watch a Colab notebook take shape. Now available for all Colab users → https://t.co/PO6OMko9qV https://t.co/ByThAqpvMJ

We are proud to partner with @deutschetelekom to make Perplexity Assistant the native feature on their new AI Phone. https://t.co/rYwsuySLW4

📲 Audio generation, on-device, off-line, for the first time! We’ve partnered with @Arm to bring generative audio to mobile devices, enabling high-quality sound effects and audio sample generation directly on-device with no internet connection required for the first time!… https://t.co/NnmaKU3To3

THIS WILL ALWAYS TAKE ME OUT https://t.co/XvK5UfTg7Y

remember, the best model depends on the use case: - chatgpt is best for large token output - gpt-4.5 is best choice for creative writing - gemini 2.0 is the best for large context input - claude performs well in large context coding input and output - chatgpt and grok are

Strongly recommend. Faggin designed the first commercial microprocessor — the Intel 4004 — and has spent the last decade+ studying and funding research into consciousness. The more I read, the more I think the materialist paradigm is on its last legs. Bring back the magic. https://t.co/ZI5MvbbELo

part of the reason this advice is somewhat unobvious is actually that a lot of these “viral apps” are low status in silicon valley, but what you don’t realize is that a lot of silicon valley was originally built by low status nerds at the start. (see more below) https://t.co/1JjyqTHCV2

great relationship with your mom 101: - talk to her everyday till she hangs up (important) - regular good mornings & good nights - create talking points with her (share mundane gossip, complain about your day) - gift her a smartwatch so you can compare step counts every day (i

"Failure is irrelevant unless it is catastrophic" Take the risk. https://t.co/Eccv0VounM

Blue Ghost’s shadow seen on the Moon’s surface! We’ll continue to share images and updates throughout our surface operations. #BGM1 https://t.co/iP7fWOSths

Just in, our #GhostRiders downlinked another incredible Moon shot following a successful touchdown! Image shows the Moon's surface and Earth on the horizon. Blue Ghost's solar panel, X-band antenna (left), and LEXI payload (right) are also in view. #BGM1 https://t.co/UIMB0ON3k3

You are too careful use this simple formula to decide whether that risk is worth taking (% chance of FAILURE ) * (FAILURE outcome) + ( % chance of SUCCESS) * (SUCCESS outcome) = good idea / bad idea Spoiler alert: it’s almost worth taking that risk


This list is automatically updated weekly (latest: 14 May 2026).