Veilance

Introducing the Veilance Intelligence Network. Search across thousands of real life website visits, and compare website behavior across multiple countries and real user sessions. Learn more today: veilance.org/intel-network Get paid to upload anonymized telemetry: veilance.org/paid-to-browse

Morning dev update The API is coming along, I’ve been doing consulting for another company related to Veilance and the Intel network which is why it’s been quiet lately. We have a plan for infrastructure for Verity and are working towards a funding round for it. This will take some time as it’s a significant advancement in the company itself.

We have paid out a total of 2,257,335 solana:H4scGEd64jRY67EcDmEK9XuMnBbkkE2tNtFM2Ni3pump

I'm going to have to automate the payouts now, there is too many for me to be able to go through them. We're currently sitting at about 20k per week.

payouts have started going out

Morning dev update API is almost done and people will be able to integrate the intel network and Verity directly into their own workflows. We are having increased user count and website activity across the board, with a total of around 100 unique wallets uploading telemetry constantly. This is significant as it shows that the product is being used across all browsers not just Chrome in general.

Morning dev update We spent most of yesterday recreating the current PHP exploit and verifying across our infrastructure that we are not vulnerable to any of the exploits. A couple of days ago we released all the training data for Verity up until 27B-v2. This is a total of around 10GB of data and you can find all of them here: huggingface.co/Veilance/data…. Verity-32B-v3 is done training and sitting in a private repo on Huggingface waiting for deployment. We are working on getting the infrastructure setup to transition from OpenAI to Verity. Yesterday with our public creation of the PHP exploit we were mentioned in a post, this post got a lot of attention and went semi viral causing us to get around 50 users overnight. Current statistics can be seen below: ~220 active chrome users ~6.8k unique website visitors ~327k requests Today we need to keep working on the API and finish that for release.

Another major milestone for Veilance and the entire ecosystem. After almost a month of development and training, we've finished six rounds of models, ranging from 4 billion to 32 billion parameters, with an infrastructure cost of $3,511.02 USD. To accomplish this training, we leveraged infrastructure hosted on lambda.ai, equipped with: - 56 vCPUs - 400 GB of RAM - 4 NVIDIA A6000 GPUs Training at this scale requires massive amounts of computational power, and thanks to the @nvidia Inception program, we leveraged our access to GPU credits to cover the infrastructure cost. This support provided us with access to the compute power needed to complete the Verity training and reach a major milestone that would have otherwise required massive amounts of out of pocket expenses we can now use for marketing, and fund raising. The result? Verity: a large language model built to specialize around privacy on the web. Unlike general use LLMs, Verity was developed to help users understand how websites collect, process, and share their information. Its capabilities include: - Analyzing websites and their privacy policies. - Providing advice and recommendations to improve online privacy. - Explaining complex privacy information in language easy to understand. - Accessing and analyzing intelligence collected through the Veilance Intelligence Network. - Helping users understand relationships between what websites disclose in their policies and what their behavior actually reveals. This is way more than just completing a training run. We've built specialized AI infrastructure around real browser telemetry and privacy intelligence, with the goal of making online privacy much more transparent, understandable, and accessible. A huge thank you to @nvidia for access to the Inception program for their help in making this possible. The completion of Verity is a massive step forward for Veilance, and we're ecstatic about what comes next.

Verity is ready for production. Verity-32B-v3 will be used as our production model and will be replacing our current OpenAI model once we have the infrastructure ready to support it the switch. To celebrate we have released more than 10GB of training data across 4 different versions that we used to train Verity. You can find the datasets in the following links: - Version 1: huggingface.co/datasets/Veil… - Version 2: huggingface.co/datasets/Veil… - Version 3: huggingface.co/datasets/Veil… - Version 4: huggingface.co/datasets/Veil… These datasets each make up a training set for the Verity AI and was used during the training of one of the models we have released.

Current Veilance statistics. - Roughly 210 active users across Chrome, Edge, and Firefox - 90,999 total telemetry uploads, 86,572 accepted uploads - Installed all across the world - 6.13k unique visitors to the website in the last 30 days - 2GB of data served in total for the last 30 days - ~285k requests in total in the last 30 days - Over 1.6m solana:H4scGEd64jRY67EcDmEK9XuMnBbkkE2tNtFM2Ni3pump paid to telemetry uploaders This is the ecosystem we are building. Users across the globe who want privacy and insight.

Morning dev update Been working on the API for integration into workflows for users. There will be a free version and paid version of the api and an SDK available. Once the API is done we are going to start working on the browser for mobile devices. This process will take some time and is a massive undertaking. Verity-32B-v3 has gotten a 99.98% token accuracy and will be used in production. We are working on getting some GPUs purchased so we can self host it. We have roughly 90,000 telemetry uploads and have paid out 1.6m solana:H4scGEd64jRY67EcDmEK9XuMnBbkkE2tNtFM2Ni3pump to uploaders in total.

Morning dev update Currently doing solana:H4scGEd64jRY67EcDmEK9XuMnBbkkE2tNtFM2Ni3pump PAYOUTS and have over 18,000 uploads to process. That's roughly 2.5k uploads per day for a week straight. Verity-32B-v3 is finished training! We are running testing and will provide more information as we find out more. The next steps for Veilance is to start working on the mobile browser, this will take some time as it's a big project. We will update you as we get the browser started and rolling out.

Morning dev update Verity-32B-v3 is training and running. This will take awhile as this is the largest model we have trained to date. We are currently at over 80,000 telemetry uploads. This is a significant amount of data that has been uploaded to us. Yesterday, we finished OpenAI SDK compatibility for Verity. Developers can (soon) access Verity directly through the API using the SDK and interact with Verity without installing dependencies or learning a new integration. This is a significant milestone and reduces integration friction while simplifying developer onboarding. It also provides a industry standard for interacting with Verity.

Verity is coming to the Veilance API with OpenAI SDK compatibility. Developers will be able to access Verity using their Veilance API key, with no custom SDK required. Coming soon.

morning dev update yesterday we released the new version on the Veilance Intel Network. This version allows anyone to go and search a URL or domain and get information back, free users get capped at 10 sessions with limited infromation, premium users get capped at 450 sessions with access to: per county, real-time ingestion, traffic flow, and more. use the new intel network here: veilance.org/search/ Verity-32B-v3 is training still, probably will be for awhile.

The newest version of the Veilance Intelligence Network is out now! There is no account required, however Premium users can view real-time sessions as they come into the network while Free users have a limited data view. Try it out now: veilance.org/search/

What does @Polymarket actually do when you cisit the site? We put it to the test using Veilance. In 30 seconds, Veilance observed: - 421 network requests with 40 to outside services - 6 thirdparty contacted domains - 4 identified trackers - 231 browser actions with 61 storage actions We compared the observed behavior against their privacy policy with Verity. Here's what Verity had to say: The policy discloses use of cookies, analytics, advertising, and trackers. These disclosures align with some observed activity. However, it is less specific about the thirdparty services involved, the browser fingerprinting techniques observed, and the extent of identification and storage activity. While the policy broadly disclosed collecting device and usage information, it doesn't communicate the level of activity measured during this session. To be clear there were no direct contradictions identified, but a few observed behaviors were only partially covered by the policy. This is why we're building the Veilance Intelligence Network. Privacy shouldn't just be what a company says. It should also be about what its website actually does. Beyond the fine print: veilance.org

78,000 uploads. 1.6m solana:H4scGEd64jRY67EcDmEK9XuMnBbkkE2tNtFM2Ni3pump paid out. We're just getting started.

The future of domain intelligence isn't blocklists. It's evidence. Track websites across different browser sessions, countries, and timelines. See observed behavior, compare activity, track changes, and investigate the evidence. The Veilance Intelligence Network is coming.

Training Verity-32B-v3 nitter.cf/i/broadcasts/1DxleVgPj…

Morning dev update We started increasing visibility into the Veilance Intelligence Network for free users. Free users will still be capped at 10 sessions but will be able to view more details then they can now. We also started building a permanent landing page for VIN where you will be able to go and search domains or URLs. This will make it easy for anyone to use it. Basically the idea is to show up and have a search bar where you enter your domain and get info back. Verity-72B-v3 is training as we speak. This version is going to take MUCH longer to train as the model is much larger than anything we have trained so far.