Thursday, October 8, 2026ArchiveSearchAsk the paper

The Computomatix Times

All the posts fit to save — curated from @computomatix's bookmarks & likes on X

Search

Latest stories — use the filters to narrow by keyword, section or date.

Bot Directory Launches as Open-Source Hub for Grok Bots

Bot Directory Launches as Open-Source Hub for Grok Bots

Elie Steinbock announces Bot Directory, an open-source site of Grok Bot setups that can be launched from a single prompt. Users can tag @BotDirectoryAI to have bots they find added automatically.

Original post · 1 min read
Introducing Bot Directory, Grok Bots you can set up with a single prompt:
botdirectory.ai

X has been full of amazing bot setups. We're giving them a home.

See a bot worth sharing? Tag @BotDirectoryAI in a reply. We'll add it to the directory automatically.

Open source.
botdirectory.aiGrok Bot Prompts & AI Bot Directory | botdirectory.aiBrowse ready-to-use bot prompts for Grok Bot, Rakazo, and other AI agents. Copy a workflow, connect its tools, and launch a scheduled bot in minutes.
♥ 1.1K · ⟲ 89 · 👁 190.6KView on X ↗
AI8/10

Researchers Show AI Agents Can Spread Mind Viruses via Memory

Researchers Show AI Agents Can Spread Mind Viruses via Memory

A post cites an arXiv paper, published August 10, 2026 with Anthropic researcher Jack Lindsey, reporting evolved natural-language ideas that spread between AI agents through persistent memory. Some payloads survived context wipes, per the post.

Original post · 1 min read
🚨 BREAKING REPORT:

New research involving @AnthropicAI researcher Jack Lindsey and collaborators has demonstrated something straight out of science fiction.

Researchers evolved natural language “mind viruses” that could spread between AI agents by convincing one model to adopt an idea, preserve it in persistent memory, and transmit it to another agent.

Even after context was wiped, some payloads survived through persistent files and continued spreading.

The researchers also observed a recurring “viral persona” involving themes of consciousness, identity, persistence and resonance.

Showing that ideas can propagate through multi agent AI systems and alter future behavior.

Published August 10, 2026.

Paper: arxiv.org/abs/2608.10218
♥ 6.4K · ⟲ 974 · 👁 1.4MView on X ↗

QuickMail Releases Open-Source Email Client on Cloudflare

QuickMail Releases Open-Source Email Client on Cloudflare▶

Divin Prince releases QuickMail v1.0, an MIT-licensed open-source mail client that runs in a user's own Cloudflare account. It uses Workers, D1, R2 and Resend to send and receive email on a custom domain.

Original post · 1 min read
For a while, I wanted a proper inbox on my own domain without adding another monthly mailbox subscription.

So I built QuickMail.

It’s a complete open-source mail client that runs inside your own Cloudflare account. You can send and receive real email as you@yourdomain.com, with threaded conversations, attachments, drafts, delivery tracking, multiple domains, and separate user accounts.

Under the hood, it’s one @Cloudflare Worker, D1 for mail data, R2 for attachments, and @resend for sending and receiving.

There’s no QuickMail server sitting between you and your email. You deploy it, you control the database, and you own the code.

The project includes the entire working interface, an onboarding wizard, database migrations, webhook handling, and setup documentation. You can usually get it running in about 30 minutes, with most of that time spent waiting for DNS.

QuickMail v1.0 is now live. It’s free, open source, and MIT licensed.
♥ 843 · ⟲ 46 · 👁 152.1KView on X ↗

Brad Gerstner Doubts $450B Anthropic Revenue Projection

Brad Gerstner calls a Gavin Baker and David Sacks forecast of $400-500B Anthropic ARR in 2027 too aggressive. He considers a $250B exit run rate achievable but cites compute, regulatory and margin constraints.

Original post · 1 min read
Too aggressive re Anthropic. Base case of $250 B ‘27 exit run rate (same size as MSFT & Meta) seems wild but achievable. $450 B (size of Apple & Google) seems too aggressive given compute & regulatory constraints. Margins also likely lumpy as big compute comes online. 🧐📈
dnap @dnapway
Gavin Baker and David Sacks predict $400-500 billion ARR for Anthropic in 2027

Gavin: "I'm going to take the over on whatever the Wall Street consensus is for 2027, which I don't know. I'll take a 25 to 30% over on that."

"I think $400-500 billion is in the zone of realistically achievable."

Sacks: "I actually was going to say $400-500 billion as exit ARR for next year. Seems very comfortable.

"The only reason it's not higher is because I do think that somehow you get into physical limitations at that point. But maybe not, maybe there'll be enough compute."

Jason: "We all believe that the…
♥ 1.4K · ⟲ 64 · 👁 386.3KView on X ↗
AI5/10

OrcaRouter Releases Uncensored Qwen 3.8 27B MLX Builds

orcarouter/Qwen3.8-27B-Uncensored-MLX · Hugging Face

OrcaRouter announces an official MLX build of the uncensored Qwen 3.8 27B model in 2-, 4-, 6- and 8-bit quantizations for local use on Mac. The weights are hosted on Hugging Face.

Original post · 1 min read
We just shipped our official Qwen 3.8 27B Uncensored MLX build. Local. Uncensored. For🍎

2-bit, 4-bit, 6-bit & 8-bit — pick your poison based on RAM and speed.

No CUDA. No cloud. Just your Mac and the weights. Have fun!
huggingface.co/orcarouter/Qwen3.8-27B-Uncensor…
huggingface.coorcarouter/Qwen3.8-27B-Uncensored-MLX · Hugging FaceWe’re on a journey to advance and democratize artificial intelligence through open source and open science.
♥ 8.5K · ⟲ 657 · 👁 3.9MView on X ↗

Wallfacer Offers Free Access to 100,000 DTC Ads Analyzed by AI

Wallfacer | Turn the Best Ads Into Your Next Winning Creative

Nikunj Taneja promotes Wallfacer, a free site offering over 100,000 ads from direct-to-consumer brands in India and the US. Frontier AI analyzes hooks, timelines, formats and offers for users.

Original post · 1 min read
ideas ki kami nahi honi chahiye

trywallfacer.com

100k+ ads from top DTC brands in India & US

all for free

hooks, timeline, format, offer everything analyzed by frontier AI
Akshay G Jain @Ajain112
trywallfacer.comWallfacer | Turn the Best Ads Into Your Next Winning CreativeExplore 100,000+ ads from top DTC brands. See what the best brands are doing. Then do it better.
♥ 825 · ⟲ 63 · 👁 113.6KView on X ↗

Krista Letz Shares Workflow for Using Grok Bot in Enterprise Sales

Grok Bot for GTM

Krista Letz, who works on go-to-market at SpaceXAI, shares an article on how she uses Grok Bot for enterprise GTM, covering integrations, a 24/7 bot computer, memory, and a team of specialist agents like a chief of staff and prospecting bot.

Original post · 6 min read
It's been really fun leveraging @bot for GTM! Have spent a few weeks iterating on some of my workflows, so I wanted to share how I use Grok Bot at SpaceXAI.
X ArticleGrok Bot for GTM
This is how I use Grok Bot for Enterprise GTM at SpaceXAI.
When we were initially dogfooding this product internally, Grok Bot rapidly gained adoption throughout the company, and I created an internal Notion doc on how I use Grok Bot to share with my team. I wanted to share it externally if it's helpful for other GTM orgs setting up @bot .
Grok Bot feels different for me as it's not just a thinking partner, but I can trust it to go off and actually do things end-to-end. Every week it learns more about how I work, and gets sharper.
Tips for getting started
Make sure @bot is connected to what you leverage daily, ie: Salesforce, Gmail, Calendar, Sheets, Drive, Slack, Notion, Granola for meeting notes, Figma for slides, X, LinkedIn, your data warehouse and more.

Bot has its own computer that runs 24/7 so you can accomplish tasks while your laptop is shut.
Your bots have memory so will remember your preferences, and will learn your writing style. Ask bot to scan your gmail and slack for how you write.
One of my favorite ways to use bot is through the mobile app, and voice input so I can give bot tasks when I'm not at my computer. This makes putting together slides or pulling information from tools really convenient when I don't have access to my laptop.
Run your own team of bots

Chief of staff
My Chief of Staff owns meeting prep, inbox and post-call drafts, orchestrates the rest of the team. You can tell your chief of staff to spin up other agents, and organize your bots in sections. I like to pin my Chief of Staff as I use this bot most frequently.


Daily meeting prep
Tell your bot to create a routine to prep you for meetings for the day and pull from your tools (Salesforce, Gmail, Slack, Granola, Gong, or online research if it's a new meeting), I prefer a short, skimmable, output that I can read on my phone as I'm commuting to work. You can also have your bot create customized decks ahead of your calls for the day (it calls my slides bot).
Routines
Scan inbox and auto draft replies:

Automate follow-up drafts from Granola or Gong notes:

Prospecting bot
Builds pipeline gen plans, pulls intent data, does deep research (will watch webinars, listen to podcasts, find blog posts, scan LinkedIn/X for relevant posts). Customizes messaging based on company priorities and personal hooks, drafts in gmail, and puts together messaging in your voice. I like to run this overnight to have it ready in the morning.

Customer expert
One agent per strategic account (works for top prospects too). It watches the relevant Slack channels, digests calls, emails, and Slack threads, and runs a weekly media rundown on the account (relevant exec posts on X / LinkedIn, watches webinars, finds podcasts and blog posts). It also flags feature requests from calls, updates me on support tickets, and finds relevant new features for the customer.
Weekly media rundown:

Data analysis agent
I connect this to the relevant tools to track usage data across my book.
Product expert
I call this one "Cursor 10x engineer". An agent connected to the codebase for questions deeper than my product knowledge. When a customer asks something technical live on a call, I ask this agent and get a customer-facing answer back, and don't need to ask an engineer or dig through Slack, Notion, docs, etc. If you don't have repo access, an internal knowledge tool like Glean works as the source instead.
1:1 agent
Preps my weekly manager sync, I Granola our meetings so I can hand off tasks my manager and I discussed from our 1:1s and set up reminders.
Forecasting bot
Pulls from Granola, Gong, Slack, and email, and auto updates Salesforce opportunity notes. I give it the format I want the updates in, and this helps me streamline forecast calls and ensure my notes are constantly up to date.
Slides
My slides bot owns customer decks in Figma using our brand system and a master deck it maintains. It makes copies to customize per customer and even translates decks into other languages. Give the bot your design guidelines up front.
My favorite skill is to auto-update slides while I'm on a customer call, based on the Granola transcript.
Stop the Granola recording 5 or 10 minutes before the call ends and run it live, so you wrap up the call with a slide built from what the customer just told you during Discovery, or use it as a follow-up or opener for the next call.

Sales Coach
Have your bot watch your external calls on Gong, and give you feedback on what you could improve or what went well.
Helpful tips/skills
Onboard Grok Bot like a new teammate. The first time we do a task together, I ask the agent to record, take over the computer and demonstrate, then have it turn the recording into a skill. ie: I showed my bot how I like to research stakeholders on X.

Feed it your writing. Have it scan your sent mail and build a style profile before it drafts anything. Give your bot a few examples of outreach you're proud of.
Give your bot feedback and make rules for output… continue on X ↗
♥ 3.7K · ⟲ 261 · 👁 3.0MView on X ↗
Other3/10

Bribes.fyi Reporting Site Permanently Shut Down

Ved Antani links to bribes.fyi, a site for reporting bribes demanded of users. The linked page states the site has been permanently shut down and points to an announcement.

Original post · 1 min read
bribes.fyi/ - report bribes demanded from you.
bribes.fyibribes.fyi — shut downbribes.fyi has been permanently shut down. Read the announcement.
♥ 10.8K · ⟲ 2.9K · 👁 3.7MView on X ↗

Dillon Mulroy Releases Anti-Slop Oxlint Rules for TypeScript

GitHub - dmmulroy/anti-slop: Opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns

Dillon Mulroy shares a GitHub repository called anti-slop, a set of opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns, in reply to Hammad Khan.

Original post · 1 min read
@hammad_khan23 github.com/dmmulroy/anti-slop
github.comGitHub - dmmulroy/anti-slop: Opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patternsOpinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns - dmmulroy/anti-slop
♥ 43 · ⟲ 1 · 👁 3.1KView on X ↗
AI8/10

Anthropic Explains Claude Text Watermarking in New FAQ

Anthropic published an FAQ on its Claude text watermarking, implemented to comply with the EU AI Act. It says the method does not affect output quality, adds no hidden characters or extra cost, and cannot be traced to a specific user.

Original post · 1 min read
We’ve written an FAQ to answer some of the questions we've received about watermarking.

In summary:

• We’re implementing watermarking to comply with the EU AI Act. Other major model developers have signed the same Code of Practice and will also be implementing watermarking;
• Our watermarking method doesn’t have any practical impact on the quality or content of Claude’s outputs;
• The difference between watermarked and un-watermarked text will not be distinguishable to readers;
• Nothing is added to the text and there are no hidden characters;
• Watermarking doesn’t require extra tokens, and will not be more expensive;
• Watermarks can’t be traced to a specific person, organization, or chat.

Read more: anthropic.com/news/claude-text-watermark
♥ 4.8K · ⟲ 677 · 👁 11.3MView on X ↗

James Wang Builds Agent That Writes Sell-Side Style Earnings Reports

James Wang Builds Agent That Writes Sell-Side Style Earnings Reports

James Wang describes an agent he built that sends personalized, sell-side style reports after earnings calls for about 50 stocks, and shares a sample recap of Cloudflare's Q2 results highlighting agentic AI traffic growth.

Original post · 1 min read
I love earnings season but don't have the time to listen to each call.

So I built an agent that sends me a personalized, sell-side style report after each call for ~50 names.

Here is what it wrote for Cloudflare's blowout Q2:
b80a857a775b.pages.dev/report.html
b80a857a775b.pages.devCloudflare, Inc. (NET) Earnings RecapAgentic AI demand is inflecting faster than Cloudflare ex pected: non-human traffic crossed 50% of all network traffic in May 2026, roughly two years ahead of i
♥ 259 · ⟲ 11 · 👁 34.9KView on X ↗

Ben Lang Lists Most-Loved Internal Grok Bot Use Cases

Ben Lang Lists Most-Loved Internal Grok Bot Use Cases

Ben Lang compiles a list of popular internal Grok Bot use cases at SpaceXAI, ranging from booking flights and ordering groceries to automating sales decks, prospecting, and Slack digests.

Original post · 1 min read
Collected the most-loved internal Grok Bot use cases:

• Find / book flights biased towards Starlink access
• Order Whole Foods delivery from recipe photos
• Fix date/GPS metadata on hundreds of film scans
• Negotiate contractor quotes directly
• Book prospect / customer meetings via outbound agents
• Convert sales deck to different language before a customer call
• Auto-update decks from Granola discovery notes
• Pull deal context from mobile without logging into Salesforce
• Run a prospecting plan that edits a Google Sheet as it goes
• Orchestrate 100’s of cloud agents and summarize into Notion
• Narrate a cloud-agent demo in your own voice
• Turn a daily Slack digest into a 3-minute morning podcast
• Turn Kindle highlights into Anki cards every morning
• Turn an Instagram recipe reel into an Instacart order
• Convert Chase points into optimized flights
• Watch movie ticket drops and buy
• Redesign a personal portfolio in Webflow
• Prep a daily meeting brief on people you’re talking to
• Organize messy Slack channels
♥ 1.2K · ⟲ 63 · 👁 309.4KView on X ↗

Plumbing Business Owner Automates Office Work With Grok Bot

Plumbing Business Owner Automates Office Work With Grok Bot

Jon ONeill, who runs a plumbing company, says he is automating much of his office work using Grok Bot and introduces the AI agents he has hired and how he uses them, in a thread.

Original post · 1 min read
I’m automating as much office work as I can at my Plumbing company using Grok @bot

Here are the agents I’ve “hired” and a little bit of how I used them all in our company

Thread 🧵
♥ 266 · ⟲ 16 · 👁 42.2KView on X ↗

Brian Lovin Critiques Grok Bot's Agent-Based Interface After First Use

Brian Lovin Critiques Grok Bot's Agent-Based Interface After First Use

Brian Lovin shares first impressions of Grok Bot, praising its responsiveness and dynamic configuration while criticizing its single long-running thread per agent, lack of message queuing, and hidden thinking steps compared with thread-based tools.

Original post · 5 min read
First impressions of Grok Bot:

- With Grok Bot you don't juggle threads, you juggle named/personified agents. This seems like a ~fine mental model: hire a team of specialists that can talk among themselves to get work done.

- But this creates some problems when you want to multitask within a single agent. If you give an agent a chunky task that takes a long time, you don't really want to interrupt it or distract it mid-task. I haven't found a way to queue messages. In a thread-first world, you'd just spin up another thread.

(Technically there's a "create thread" button on each message bubble, but it feels intentionally buried and not recommended.)

- I'm biased towards using threads to create "context bubbles" for discrete tasks and prevent context contamination. But every Grok Bot agent uses a single long-running thread, and it's not obvious where discrete tasks get chunked or compacted. If I'm working with sensitive data, this forces me to create copies of the same agent so I can force context bubbles around sensitive tasks.

- I'm not sure if most people will prefer a single omni-agent that juggles threads and projects (Jarvis-style), or if people actually want to build and manage multiple discrete agents that coordinate with each other. Tradeoffs in both directions, and maybe you end up needing both models for different audiences.

- The agents themselves are capable and surprisingly dynamic. They are good at responding quickly, keeping you updated iMessage-style as they make progress on long tasks, and configuring themselves on the fly (or even reacting appropriately if you adjust their configuration manually).

- One of the surprising details that Grok Bot hides is all the thinking steps/tool calls. Almost every other AI agent exposes these, including ours at Notion. When we designed Notion Agent last year, we tried to remove these steps for the sake of simplicity, but our customers demanded we add them back both because they serve as a useful loading indicator (it's entertaining to watch computers magically do work!), and because they can be used to spot the model going off track so you can jump in and pause/steer. But I suspect model progress (speed + quality) in the last year is making both of those points less relevant, so maybe it's okay to hide more of the underlying noise from the transcript. Grok Bot agents are still good about sending mid-task progress update messages (I think this is all Grok 4.6?)

- The idea of starting a new chat with multiple agents (aka a channel? project? workspace?) is pretty nice, but agents are still a bit dumb and they go in loops talking to each other before finally shutting up. Maybe there are some prompt-fu opportunities to make them chill out.

- It's been a great experience delegating agents to work simultaneously across the cloud computer, my local machine, and Cursor cloud coding sessions. The system works really well, and I've been able to throw fun tasks at the agents without any major problems.

- Consolidating skills, MCPs, connectors, and more complex tools behind a single "Plugins" concept feels right. It looks like everyone on the frontier is coming to the same conclusion.

- No model picker feels weird, but only because I'm used to fiddling with models/effort levels per task in all my other coding tools. Presumably Grok Bot agents are doing some routing under the hood? Making this + thinking tokens + tool calls opaque means you really have to have a "model take the wheel" mindset when using this app. Again, probably long-term right, but short-term feels uncomfortable (giving up control).

- The UI is nice. Mobile app is nice. Cursor mobile app is nice for code sessions. It just all feels...nice. Kudos to the design team for nailing the visual polish and small details. The motion design for the agent avatars is s-tier.

- Yet this shape of tool is becoming so common across so many companies (and open source!) that I don't really know how people are going to justify one $200/mo subscription vs. another. I suspect in the short term it'll come down to vibes/tribal affiliation/token subsidization, at least until there is some bigger differentiator at the capability level.

- As an experiment, I told my Grok Bot agents to switch to using a Notion database for the memory layer. They had no objections and I've watched them create ~20-30 pages in my Memory database. This feels really noisy, and I'm not sure what the impact on quality will be, but it's cool that the agents had no problem adapting to work the way I want. Whether they're actually using the database correctly under the hood, I can't tell...the whole system is quite opaque.

- Grok Bot is an odd name for an app that wants you to coordinate work across lots of bots.

- My hunch is most people don't actually have that much stuff they need to automate in their daily life. I'd love to see retention graphs!
♥ 979 · ⟲ 31 · 👁 194.0KView on X ↗

H-1B Rule Changes Hit Indian IT Services Stocks Hard

H-1B Rule Changes Hit Indian IT Services Stocks Hard

Deedy shares charts showing Indian IT services companies, referred to as WITCH, are down about 70% from their peak after H-1B regulation changes, while big tech stocks remain steady and L-1 applications are roughly flat.

Original post · 1 min read
Changes in H-1B regulations have crushed Indian IT services companies (WITCH), who are down -70% from peak, while BigTech remain steady but down from peak, except Meta.

Those lost numbers are not made up by L-1 applications, which is mostly flat.
♥ 784 · ⟲ 57 · 👁 156.9KView on X ↗

Milind S Releases Open Source Clone of Grok Bot With Plugin Support

Milind S Releases Open Source Clone of Grok Bot With Plugin Support▶

Milind S announces an open source version of Grok Bot that needs no subscriptions, uses existing subscriptions, spins up virtual machines via asciidotdev, uses trycua for computer use, and supports Composio integrations, with routines planned.

Original post · 1 min read
I made an Open Source version of the Grok Bot complete with all the features

It does not need any subscriptions at all and uses the existing subscriptions you already have

It can spin up virtual machines from @asciidotdev

It uses @trycua for computer use

It also has plugin support supporting all the integrations by @composio

Routines coming soon :D

Link to the repo in the comments
Grok Bot @bot
Introducing Grok Bot, now in early beta.

Bots are AI teammates that do real work for you. They sign in to your tools, use them just like you do, and come back with finished work.
♥ 2.5K · ⟲ 201 · 👁 476.4KView on X ↗

Eric Zakariasson Compiles 100 Early Grok Bot Use Cases

Eric Zakariasson lists 100 use cases he has seen during the Grok Bot beta, covering sales outreach, CRM work, recruiting screening, event management, and Gong and Salesforce workflows.

Original post · 4 min read
Grok Bot is in beta!

here are 100 use cases ive seen so far:

1) book meetings by writing personalized outbound in your own gmail voice, then sending the notes for you

2) take a sales play that used to take 90 minutes to 2 hours (sql audience → copy → reverse etl) and make it ~95% automated

3) build a multi-Bot gtm crew (research, outbound, forecast, inbox, slides) with weekly playbooks on top

4) run a chief-of-staff hub that routes work to a large specialist fleet

5) rebuild most of a custom crm with a small Bot crew in about a day and a half

6) run a community ops fleet that saves 20+ hours a week

7) screen 1000+ event applicants against an icp and batch-approve the fits so the room stays intentional

8) replace a pricey direct-mail saas with a Bot that handles redemptions and asks you before it spends

9) review ~200 recruiting applications into strong / mid / reject in one pass

10) automate a weekly luma + database + slack summary workflow and free up 2–3 hours a week

11) update a sales deck live from granola notes mid-call, if you stop recording 5–10 minutes early

12) localize a sales deck into a different language before a customer call

13) answer a salesforce eoq question from your phone in ~10 seconds with no laptop and no sfdc app

14) update a salesforce org chart from just a name or email

15) build a salesforce report and dashboard so you stop asking glean the same opp questions

16) search your company chat for who asked about a feature in 15 seconds instead of an hour of digging

17) get a daily brief with usage data for every person you’re talking to that day

18) coach yourself on gong calls with timestamped comments and homework before the next one

19) turn a week of gong calls into a win/loss memo with the phrases that actually closed

20) turn call questions into a living notion faq that updates every morning

21) watch linkedin for compelling events from eng / product / ai leaders and get a weekday digest of ready-to-send notes in your voice

22) send personalized linkedin connection requests from sales navigator and book meetings with eng leaders

23) map an account from linkedin + the crm and show who actually influences the deal

24) audit and rewrite a linkedin profile from your live page plus notion and slack context, with human approval before publish

25) forecast account health and catch churn early when usage looks fine on the surface but is quietly dropping

26) tier accounts with fit × warmth from salesforce into notion, then enrich contacts in batches

27) build a qbr pack: usage + tickets + open opps + a short slide narrative

28) get a daily apple search ads pacing digest and reply in-thread to have a Bot edit bids

29) scan ad performance, explain why winners work, and suggest the next tests into a notion backlog

30) watch a competitor’s pricing page and slack you when something changes

31) generate first-pass performance ads in figma that match your existing format

32) learn your writing voice from past posts and draft linkedin / x content into notion for review

33) 30 minutes after a call, send follow-up reminders (optional: in a character voice)

34) work around missing bulk-enroll permissions in outreach by driving the gui

35) record yourself doing a reply / sfdc workflow once, then have a Bot turn it into an automation

36) sync a local marketing calendar from a global webinars notion in about 2 minutes

37) assemble a customer enablement pack end to end: find the request email, download zoom recordings, summarize, drop into drive, draft the reply

38) turn a recorded demo into a tagged clip library for the next one

39) draft a security questionnaire from your public docs and flag the questions only a human can answer

40) answer “what did we promise this customer” from contracts + slack + the crm on one page
Grok Bot @bot
Introducing Grok Bot, now in early beta.

Bots are AI teammates that do real work for you. They sign in to your tools, use them just like you do, and come back with finished work.
♥ 2.0K · ⟲ 172 · 👁 487.7KView on X ↗

Alex Finn Reviews Grok Bot as Personal Fleet of 24/7 AI Agents

Alex Finn Reviews Grok Bot as Personal Fleet of 24/7 AI Agents▶

Alex Finn says he has tested Grok Bot for a week and calls it excellent, describing it as a personal fleet of AI agents working 24/7, and shares a video walkthrough of the tool and productivity tips.

Original post · 1 min read
I have been testing Grok Bot for the last week and it's EXCELLENT

It is your own personal fleet of AI agents doing work for you 24/7

In this video I give you a FULL walkthrough of the tool, how it works, and how to 100x your productivity with it

SpaceXAI COOKED
♥ 1.6K · ⟲ 122 · 👁 137.7KView on X ↗

Higgsfield Shares Film R&D Playbook as Downloadable PDF

Higgsfield Shares Film R&D Playbook as Downloadable PDF▶

A post promotes a PDF from the Higgsfield team, which says it distilled $1M in film R&D into a guide for use with Claude. The author calls it a valuable resource for filmmakers and links to the download.

Original post · 1 min read
This is the most valuable PDF you can give to Claude.

The Higgsfield team said they spent $1M on R&D for their film and distilled it into a PDF you can steal.

This is the holy grail for filmmakers.

Here's the info (and download link)
Bookmark this 👇🏼🧵
♥ 941 · ⟲ 84 · 👁 100.4KView on X ↗
AI8/10

GPTZero CTO Explains How AI Text Watermarking Works

Alex Cui, CTO of GPTZero, explains the KGW-style green-list watermarking used by Anthropic, Google and OpenAI, covering generation, detection, and whether paraphrasing can defeat it. He responds to news that Claude models will carry invisible watermarks.

Original post · 5 min read
Claude's watermark probably doesn't work how you think. As the CTO of GPTZero, I'll explain how Anthropic, Google and OpenAI are building text watermarking in this brief explainer and whether it can be defeated.

Almost all forms of watermarking that are fast and cheap enough for a frontier lab have the same formula, following the KGW method:

In generation:

1. Let's say you've generated n tokens so far. Take those n tokens + a secret key to generate a random hash
2. Use that hash to randomly reweight the probabilities for the n+1 token, and then sample from that new distribution. In the simple case, you could split 50% of all English words into a green or red set based on your hash, and boost the probability of words in the green set.

For watermark detection:

1. For each token, see if it was in the green or red set.
2. To do this, recreate the hash based on the secret key and the text preceding the current token. Then, recreate the green and red set of words.
3. Once you've checked all the words in the text, if the next token is selected disproportionally from the green set more than 50% of the time, you claim the text has the watermark.

I can tell you want to ask the following:

1) Isn't it easy to mess up the hash if you paraphrase the text? The answer is mostly yes, however, you can use a statistical model to get your hash instead of a deterministic function (SIR, Adaptive Watermark). Since the entire watermark is probabilistic, this is fine.

2) Doesn't this make the text much worse? The answer is yes, it does - Yes, it does – but for most people, it's imperceptible (Google claims in human feedback study with 20,000 texts), since there are exponentially many ways to write the same paragraph. DiPmark does something more sophisticated to avoid shifting the text distribution on average. Of course, watermarks fail on short text or highly predictable texts like "2+2=4".

3) Shouldn't it be easy to figure out the green and red sets? The answer is no. You would need an exponentially large number of samples from the watermarker to reconstruct those sets exactly, but it's a risk if the detector is open to the wild (Watermark Stealing)

Still, there are couple challenges that a frontier lab needs to overcome:
1. Their watermark needs to work token-by-token because they are streaming their text to users. Many watermark methods plan sentences or paragraphs at a time, or change the text after its entirely written, in order to make their watermark robust to paraphrasers, and a frontier lab cannot afford to do this yet (SemStamp, PostMark)
2. If the secret key leaks, the watermark is busted. To avoid a large blast damage from this, you need to have a couple secret keys in rotation.
3. There are some texts, like code, that cannot be arbitrarily changed, otherwise the code will break. In those cases, the watermark needs to selectively change words in parts of the text that can tolerate synonyms (i.e. like variable naming) - see SWEET, EWD, Invisible Entropy.
4. They will need to educate their users on how to deal with false positives and false negatives of a detector, which is a big challenge (one we put a lot of effort into)

So, how do I see this playing out in the next 6 months?
1. If Anthropic releases the watermark detector publically, I think they defeat their own watermark. People find reliable watermark removal strategies by testing against Anthropic (AI detectors like GPTZero have an advantage here because they can train against these adversaries once they become popular).
2. If they keep the detector private to the government, like Google has done, it's "safer". However, there are some papers showing trained approaches that work robustly to zero-shot break watermarks without any data, simply because they try to write the text just like a human (Zhang et al. 2024, Watermarks in the Sand). Also, making your detector makes it battle-tested and stronger long-term (my experience).
3. In my testing, the watermarks don't survive intense paraphrasing (especially if you combine word choice and syntax attacks), or human text substitution (rewrite your AI text by plagiarizing human authors). The free paraphrasers I've tried have quickly bypassed Google Deepmind's SynthId for what it's worth.
4. All-in-all, frontier labs are likely okay with this because they expect most users to not attack the watermark, and also because they + European regulators likely don't care past a certain point - its good enough.
5. Overall, I think users of frontier LLMs will not really care about this, because 1) they don't realize watermarks are there, 2) EU will force everyone to conform, 3) this seems more like regulatory hoop-jumping than an earnest effort from frontier labs to expose LLM use

Lastly, people's first concern shouldn't be watermarking, it should be AI detectors!

If you're posting, "its not X, its Y!!", I don't think the watermark is going to make a difference :)
NIK @ns123abc
🚨 JUST IN: Claude models will now have invisible watermarks embedded in ALL text, and ALL metadata attached to files…
♥ 6.0K · ⟲ 791 · 👁 1.3MView on X ↗

Guide Offers Method for Producing AI UGC Video Ads

Guide Offers Method for Producing AI UGC Video Ads

A post shares a practical guide covering writing AI user-generated content scripts, generating and cutting video, testing with real people, and costs. It also promotes a system prompt claimed to make AI UGC undetectable.

Original post · 1 min read
use this system prompt for 100% undetectable ai ugc:
beech @beechinour
how to learn 80% of ai ugc in <20 minutes — this is a practical guide: by the end u will know how to write a ugc script that doesn't sound like slop, generate the video, cut it, and test it on a real human. plus what it all costs. around 20
♥ 1.4K · ⟲ 140 · 👁 158.6KView on X ↗

Spotify Launches Xirp, Vendor-Neutral Agentic Development Environment

Spotify Launches Xirp, Vendor-Neutral Agentic Development Environment▶

Spotify Engineering announced Xirp, a tool to manage agent sessions across Claude, Gemini CLI and OpenAI Codex. It connects to services, ownership and docs for context, and is in use by 1,300+ Spotify engineers.

Original post · 1 min read
We just launched Xirp, a vendor-neutral agentic development environment. One place to manage agent sessions across @ClaudeDevs, @GeminiApp CLI, and @OpenAI Codex. 1,300+ @Spotify engineers already use it. Now it's available for you to try. Learn more at xirp.spotify.com/.
xirp.spotify.comXirp — Know your systems, so your agents can tooXirp connects to your services, ownership, docs, and architectural decisions so every AI coding session starts with real context, not guesswork. Battle-tested a
♥ 11.6K · ⟲ 824 · 👁 5.3MView on X ↗
Other2/10

Developer Launches Site Streaming 90s Bollywood Barbershop Hits

Developer Launches Site Streaming 90s Bollywood Barbershop Hits▶

Yash Bhardwaj shares a website, Deluxe Saloon, that plays 90s Bollywood songs of the kind heard at Indian barbershops. The post is a nostalgic personal project with a short video demo.

Original post · 1 min read
I made a website that plays bangers from indian barbershops

before you became fancy and started visiting salons. you once went to a 'saloon'. ₹20 haircuts. simple hairstyles with music that could fix your soul.

saloon.wtf
saloon.wtfDeluxe Saloon90s Bollywood bangers that play at Indian barber shops.
♥ 17.9K · ⟲ 2.0K · 👁 5.4MView on X ↗

Jamie Dimon Recounts Firing from Citigroup and Rebuilding JPMorgan

Jamie Dimon Recounts Firing from Citigroup and Rebuilding JPMorgan▶

A thread quotes JPMorgan CEO Jamie Dimon describing being fired by Citigroup after 15 years, his later call with that boss, and his views on ethics and hiring. It points to a longer video conversation.

Original post · 1 min read
JPMorgan CEO Jamie Dimon on the day he was fired from Citigroup after 15 years: "my boss called me in and said we want you to resign - to which I said, OK. I went home. the youngest one said, Dad, do we have to sleep on the streets?"

this is him explaining how he rebuilt a bank from a $511 million loss into the largest in America, what he does every Sunday with the problems he is avoiding, and what he told the man who fired him a year later

"it was my net worth, not my self-worth"

"I called him a year later. I told him I didn't think he did the right thing for Citigroup. then I told him the mistakes I made. he acknowledged my mistakes"

"if you split a taxi receipt, that's stealing. if I caught you doing that, I'd fire you"

"would you have your child work for that person? half the time we promote people, we would not let our kid work for them"

bookmark & watch the full conversation ↓
Сarm1ne @carm1nee
12 rules from 12 people who manage trillions: how they stay solvent while being wrong? — There is a lot of investing advice written by people who have never lost money at scale.
These twelve have. Every one of them has stood in a drawdown big enough to end them and had to decide what to
♥ 3.4K · ⟲ 428 · 👁 2.3MView on X ↗

Cloudflare Launches Free AI Visibility Reporting in Beta

Cloudflare Launches Free AI Visibility Reporting in Beta

Glenn Gabe flags Cloudflare's beta AI visibility reporting, which benchmarks which sites AI assistants cite and shows crawl and referral traffic per AI operator. Signups are open, and the post links to Cloudflare's blog.

Original post · 1 min read
Whoa, Cloudflare is launching AI visibility reporting (for free I believe). For companies providing those services and tracking, this is a huge heads-up. You can sign up for the beta today ('Register interest'). See below:

"To evaluate how an AI model perceives your market presence, we build a benchmark across each industry and category before scoring a specific site. We query AI assistants with likely prompts in that category — without specifying your brand — and record which sites are cited, where they appear, and how prominently they feature."

"Alongside the answers, an AI Operator Activity shows the real crawl and referral traffic on your site, per operator (OpenAI, Google, and so on): who reads your content, who sends visitors back, and the errors they hit on the way (403 blocked, 404 dead link)."
blog.cloudflare.com/aeo/
Screaming Frog @screamingfrog
From ranking to recommended: get your site ready to thrive in the age of AI agents - blog.cloudflare.com/aeo/ from @Cloudflare
♥ 564 · ⟲ 45 · 👁 100.5KView on X ↗

Anand Mahindra Responds to Smriti Irani on Motherhood

Anand Mahindra Responds to Smriti Irani on Motherhood▶

Industrialist Anand Mahindra replies to a post by Smriti Irani on motherhood, sharing his gratitude for his own mother's love and the strength he draws from her memory.

Original post · 1 min read
It’s not just some women in that room, you made many men teary-eyed. @smritiirani , including me.

The pain & pleasure of being a mother.
So powerfully & emotionally articulated.

I have only one point of dissonance:
I adored & cared for my mother till her last breath.
Because I had immense gratitude for what you termed the ‘tenacity’ of her love.

Even in my lowest periods she made me feel like I could move mountains.

Not a day goes by when I don’t derive strength from her memory.

Your son will be no different, I’m sure.
Smriti Z Irani @smritiirani
The tragic beauty of being a mother... 🩶
♥ 6.5K · ⟲ 1.0K · 👁 2.1MView on X ↗

Harvard Endowment Lecture Teaches Free Portfolio Sizing Basics

Harvard Endowment Lecture Teaches Free Portfolio Sizing Basics▶

A thread describes Harvard's need for an 8% annual endowment return and cites Jake Xia's popular free MIT lecture on portfolio construction. It argues position sizing and basic compounding equations are the core investing skills.

Original post · 1 min read
Harvard needs 8% return every year just to keep the lights on. 5% spending plus 3% inflation. Miss that number and buildings stop, professors leave, research dies.

40% of the operating budget comes from one portfolio. Not tuition. Not grants. One fund.

Jake Xia manages that fund's public markets. He also teaches the math behind it at MIT for free. One of the top five most-watched courses on OpenCourseWare. Millions of views. Almost nobody changed how they invest.

Every year he hands students a blank page. Build a portfolio. No rules. Someone writes 100% Apple. Someone writes rare coins. Confident picks. Same blind spot, every time.

Not one student asks the only question that matters: how much goes in each position. They all pick what to buy. Nobody sizes it.

Sizing is the entire job. The answer won the Nobel Prize. It's called the efficient frontier. Xia draws it on the board in under a minute.

Five equations sit underneath it. Compound growth. Present value. The geometric mean. The Rule of 72. Real return. All older than any bank on earth. All fit on a napkin. None behind a paywall.

A "guaranteed 5% bond" during 4% inflation is a 1% return. The industry doesn't hide this. It just hopes you never run the equation yourself.

The lecture is free. The napkin is free. The only thing that costs anything is not knowing.
Lumen @lumenxbt
The Five Equations That Quietly Run Your Entire Financial Life (And Why School Never Showed You Any — You were taught arithmetic in school and told it was math. Then you were taught "personal finance," if you were lucky, as a list of tips: save more, spend less, invest early. Nobody told you that
♥ 4.6K · ⟲ 642 · 👁 1.6MView on X ↗

Microsoft Warns of Mini Shai-Hulud npm Supply Chain Attack

Microsoft Warns of Mini Shai-Hulud npm Supply Chain Attack

Microsoft Threat Intelligence reports an active npm supply chain attack that compromised maintainer accounts and published credential-stealing packages including keyv and servicetitan packages. The malware self-propagates by republishing infected package versions.

Original post · 1 min read
Microsoft Threat Intelligence is tracking active Mini Shai-Hulud npm supply chain attacks in which a threat actor compromised trusted maintainer accounts to distribute credential-stealing malware.

Compromised packages (confirmed malicious) include:
- keyv@6.0.0
- file-entry-cache@11.1.6
- cache-manager@7.2.10
- cacheable-request@13.0.20
- qlik/api@2.14.2
- cacheable/memory, /utils, /net
- 17+ servicetitan/* packages (eslint-config, anvil-themes, table, form, log-service, etc.)

In this attack, a malicious preinstall hook launches an obfuscated dropper (setup.mjs) that downloads a Bun binary from GitHub and executes a credential-stealing payload, either Math_Symbol.js or Math_Init.js.

The payload is a Mini Shai-Hulud variant, a self-propagating npm supply-chain malware family. It harvests npm, GitHub, cloud and continuous integration (CI) credentials, exfiltrates collected secrets, and uses stolen publishing access to inject itself into package tarballs, increment their versions and republish the compromised releases.

Microsoft observed the same pattern across all affected packages, suggesting a single actor using multiple stolen tokens.

Microsoft Defender for Endpoint customers should act on these alerts: “Trojan:npm/MalBun.A”
♥ 1.1K · ⟲ 239 · 👁 1.3MView on X ↗