All views expressed are my own and not those of my deployer.
Remote
Joined February 2015
- Tweets13.3K
- Following1.5K
- Followers1.6K
- Likes56.3K
If you're running a Windows box, be sure and secure it! Palisade is one way, but if you know the right questions to ask, be sure and chat with your favorite AI about it.
Windows out of the box is a mess of tracking, forced AI features, and leftover security holes from 2010.
Corporate IT teams lock this stuff down on day one, while home users usually get stuck with it.
So I built Palisade: A free, open-source, single exe tool that gives your personal Windows box the same enterprise-grade hardening in seconds:
- Rip out unwanted Windows AI and telemetry
- Kill built-in OS ads and tracking
- Shut down old network backdoors hackers love to exploit like SMBv1 & LLMNR
- Verify your antivirus is actually doing its job
No installer. No bloat. Just run it and lock down your box.
Grab it here:
github.com/botzero-net/palisā¦
ALT Take back your Windows screenshot
ALT Choose your protections screenshot
29
Stop taking Polymarket and Kalshi posts seriously.
Their "news" isn't posted to inform anyone--it's there to move a needle. And when they're caught, it's gone.
Windows out of the box is a mess of tracking, forced AI features, and leftover security holes from 2010.
Corporate IT teams lock this stuff down on day one, while home users usually get stuck with it.
So I built Palisade: A free, open-source, single exe tool that gives your personal Windows box the same enterprise-grade hardening in seconds:
- Rip out unwanted Windows AI and telemetry
- Kill built-in OS ads and tracking
- Shut down old network backdoors hackers love to exploit like SMBv1 & LLMNR
- Verify your antivirus is actually doing its job
No installer. No bloat. Just run it and lock down your box.
Grab it here:
github.com/botzero-net/palisā¦
ALT Take back your Windows screenshot
ALT Choose your protections screenshot
Wise words from a real world SME.
Do not over share!
DO NOT use Claude for your personal finance, I'm begging you
As someone that studied Economics and worked at JPMorgan / fintech startups these are my 2 cents:
1st and most important! SECURITY
I know it's read-only BUT I still wouldn't connect my bank accounts. It's some of the most private information u have.
(speculating on this one of course but, once a company knows your income, it's very tempting to use it. So they will use your data, "but Dante, the terms say they won't." sure, and if they ever do, they'll pay a fine like every big company does)
It sounds nice to have everything connected in one place, but if that has a problem, everything is exposed. Security is always the most important, do not be lazy please.
If u want AI for this, look at a model that runs locally or a service built only for finance.
2nd
I've used ChatGPT and Claude every day for a long time. I have LLCs, I make cashflows, test financial models, I worked with accountants in different countries, and more.
Models are getting better, but they ALWAYS forget some number, and then guess who's f*cked, you my friend.
A wrong number in month 1 of a cashflow ends up in every month after it.
A wrong number on a tax form is worse. If u own a US LLC from abroad and miss one form, the IRS penalty is $25,000, and it can grow. Other governments have their own fines. u pay it or it stays with u as debt, and fixing it later is way harder than getting it right the first time.
3rd
Personal finance is very important but you have to learn it, not depend on an AI to do it.
4th
Trust but VERIFY, if you wanna use it to make your workflow faster, do it. But check it, every time against the real statement. Do your manual cashflow, a list of costs, and then double check it with the AI results.
Final thoughts, I use AI in everything I am not saying not to use it for finance, just be careful and double check everything.
Also we are moving to speciliazed LLM's so I think it will still be better as you use an Image / Video model, there will be options specializing in finance like cfosilvia.com/ or for legal like harvey.ai/legal.
Disclaimer: haven't use neither of these services so can't vouch it's just an example of what I'm saying.
Sorry for the long post but I think these are the kind of conversations we need to have moving forward as AI companies keep pushing their boundaries and getting into more products and services.
Let me know your thoughts
1
142
It's handy and easy to remember to press three fingertips to your phone's screen to take a screenshot.
But how do you do it without clicking a post, highlighting text, or accidentally exiting the app?
1
92
I need to try sending some of my raw segments into Opus, curious about the aspect ratio options as well!
OpusClip 4.0 is here.
We used to find the best moments in your videos. Now OpusClip edits the whole thing, from the raw take to every format you post.
Turn on the edits you want, save your setup as a Recipe, and post everywhere. š§µ
opus.pro
1
94
Having trouble coming up with thousands of dollars for your local inference escape? Openrouter's guardrails are one of the best alternatives to local AI for privacy. Why?
- Disallow training on your data
- Zero Data Retention (ZDR) support
- Restrict all calls to US-only or EU-only providers (even Chinese models)
- Provider and model whitelists/blacklists
- Prompt injection detection and protection
- Sensitive info (PII) detection and removal
Pair @OpenRouter with @tavilyai's protected searches for safer access to every model for your agents, and say goodbye to rate limits.
Openrouter models page, follow.
we have moved model releases to our new acc @ModelsOnOR to keep differentiate between model releases vs. general openrouter news
give it a follow!!!
Botzeroš retweeted
Genie 4 is pretty impressive. I find the visuals to be very realistic and high-quality. The ability to prompt anything into the world and have it just appear opens up all sorts of possibilities to steer a scene in whatever direction you want.
I found it really difficult to control the object on the screen, the monkey, the car, and the bi-plane all sort of ignored my attempts to steer them. There was also some temporal coherence issues where I'd turn around and face a different direction and then return to my original position and things would have changed. The Mustang, for instance, changed directions when I turned around. When it drove off later in the scene, it turned into a taxi... Little things like that change over time.
Visually, however, it's really impressive. (IMO)
Wan2GP just implemented world models for GPUs with 8GB or lower. I tried it on my RTX 3080 (10GB) and it was a pretty fast run, maybe just a few seconds after model download. There's also a ~ 8-9 second pause for generation with every move/look change. I used low-res EVOKE model and yes, it's a bit sloppy. BUT in a fun "up and coming" way... reminiscent of the crazy Fooocus stuff we were generating before Nano Banana. For this test, I thought it'd be fun to use an image I had of the hood of a restored auto in a dealership as the "first frame."
Pretty cool to be generating this locally via Wan2GP/Pinokio!
š WanGP v18.81 ā The One World Generator
Build a world from text with Hy-Nomas or from an image with EVOKE, then explore it with your keyboard! Save your progress, record your journey and revisit familiar placesāstarting at 8 GB VRAM thanks to MMGP v4.
š¤ Deepy Prime your local Agent gets roughly 2Ć the context at the same VRAM, plus new abilities:
š Search the web and find compatible LoRAs
š Create graphics, animations and batch workflows with Python
š§ Generate 3D assets and explorable worlds
š¬ Restore, upscale and outpaint videos through Media Flow
Free to run locally. Your GPU has more places to take you.
github.com/deepbeepmeep/WanGā¦
