my theory is that startups want brand people but don’t want to admit they need brand people, so they invent increasingly abstract job titles. meanwhile, the people they’re looking for aren’t applying to be a “narrative fairy”
So it looks like not only are these interview requests a phishing attempt, but the account itself is a fully fake reporter account, using a blue check to spam dozens of tech ppl
Be careful out there, folks!
https://t.co/wY1lxU6lX4
Excited to open source Sightmap!
Sightmap is a YAML spec + CLI that keeps a runtime map of your app. Views, components, requests, and memories, designed for agents to author and maintain. Every verification run gets smarter and burns fewer tokens.
https://t.co/d6Y7qeQyxj
What's the difference between good ol' human Session Replay (HSR) and Agentic Session Review (ASR)? HSR is pretty much Netflix. Hit play and watch. Rewind if you miss something.
ASR is different. Agents don't hit play. They ask a question, get a map of the whole visit, then zoom into just the part that answers it. No rewinding through 42 seconds to find the one moment that mattered.
In this post, @who_cayres and @jgw spell it out better than I ever could. Enjoy.
https://t.co/fByLti5DHV
If you have built a digital product, you've probably watched a session replay. Turns out agents can do it _much_ better than humans.
@who_cayres and @jgw break down what makes it possible for agents to review sessions at scale, without burning through all your tokens. https://t.co/sNWSozRQaL
Just worked w/ @CNN on a piece about why I wouldn't upload full financial docs (tax docs, statements, etc) to AI tools due to leakage & hacking risk.
I don't recommend connecting bank accounts to AI tools. It becomes a 1 stop shop for attackers looking to drain your accounts🤖💸
Now for spotting AI voice clones in phone calls or robocalls:
- Celeb or politician impersonation: if you receive a call from a “politician” about voting/not voting/actions to take on Election Day that’s a good sign it’s a voice clone deepfake.
- Verbose: live AI voice clones hooked up to LLMs are typically overly verbose, they speak for way longer than normal humans do
- Long pauses: someone using an AI voice clone either has to deal with LLM latency causing long pauses if you interrupt or they have to attempt to use a voice clone tool to type their replies…both cause long awkward delays and pauses.
- Pretending to be a boss/someone who can tell you what to do: Does the slightly robotic voice telling you to wire money sound like your CFO? Verify authenticity of the request first.
Now let’s talk thru spotting AI deepfake videos:
- Glitchy mouth and teeth: do the mouth and teeth fall out of sync? Do you see large teeth when a mouth is closed for a split second?
- Lip sync issues: do the words match up with the mouth movements or are they out of sync?
- Blinking in slow motion, oddly or backwards: does it almost look like the person blinking is in rewind/blinking in reverse? Does the amount of blinking increase throughout the video significantly?
- Ragged breathing movements: does the person in the video look like they aren’t breathing in a natural cadence, that sometimes they go a while without taking a breath and then their chest rises dramatically multiple times to breathe?
Let’s start with how to spot AI generated photos in September 2024 (will change in the future), ask yourself:
- Shockingly unusual: Is the pic showing surprising actions for celebrities, politicians, or cultures?
- Body parts and clothes a bit off: Are body parts merged together? In the background of the pic, are there people without faces? Are the people wearing mismatched earrings or have jewelry like earrings embedded in skin (like in their forehead)?
- Airbrushed and Saturated: Does the picture look highly saturated, with airbrushing effects around edges? Is it somehow lit from all sides at once? Are there more colors than exist normally in a typical photo?
- Looking for Pity: is the photo an airbrushed picture of a child or solider holding up a sign asking for support, money, wishes, or likes/follows? Does it have incorrect spelling in odd ways?
Tired of sifting through data so boring you find yourself falling asleep at your desk? 🥱
Cancel that 3rd coffee order, because we’re striving to bring a little more color into your data… and your day 🎨☀️
Learn more at: https://t.co/yGhtfoBYBq
@charleseblack Chatham House rules is generally used in conference halls, meetings, etc. where you're not speaking directly to a reporter but to everyone (including non-reporters) in the room and you want to make sure they don't attribute the info to you if they pass it along elsewhere.
Say goodbye to incomplete customer profiles and outdated insights 👋 One thing we love about Data Direct? The ability to sync clean, structured, AI-ready behavioral data to your destination of choice. 🚀
Get all the details on Data Direct 👉 https://t.co/Bj943fAdqA