Cadenza reposted this
AMD does it again! We love to see AMD growing presence in the AI and data center market 🚀 #AMD #TensorWave #AI #CloudCompute
Since 2018, the Cadenza team has managed investment vehicles with a focus on transformative and decentralized technologies. Some of our noteworthy investments include FalconX, Together.AI, and CoinDCX.
External link for Cadenza
New York, NY, US
San Francisco, CA, US
Cadenza reposted this
AMD does it again! We love to see AMD growing presence in the AI and data center market 🚀 #AMD #TensorWave #AI #CloudCompute
Cadenza reposted this
I'm excited to share the latest updates from Jericho Security!! #cybersecurity #features
🚀 Feature Forward Friday: Jericho's Latest Breakthroughs and Product Updates📢 Hey everyone! I’m thrilled to share our latest features and improvements from the Jericho Security product team. We’re excited to announce especially that our SMS Phishing capabilities have taken a great leap forward! Here are the top 10 highlights: 1. SMS Phishing Enhancements: Introducing new features including SMS Review Step Page, linking SMS Campaigns to Twilio, and SMS as plain text for improved phishing simulations. 2. SCIM User Deactivations: Expanded archive capability for user deactivations, crucial for clients using directory integrations. 3. Link Detonation Defense: Advanced methods to combat link detonations, including a new click tracking methodology and a honeypot system. 4. Async Processing for SMS: Improved efficiency with asynchronous processing for SMS generation and sending. 5. SCIM Group Optimizations: Performance enhancements for managing large groups, benefiting our clients. 6. Defensive measures for SCIM performance and group syncing fixes. 7. Daily Email Reminders: Feature flag option for daily email reminders for incomplete training, aiding compliance for clients. 8. Event Metrics Improvements: General improvements for events and recipient metrics and reports, offering better insights for clients. 9. Bot Link Detonation Detection: Enhanced bot detection methods, including a minimum click time threshold and invisible honeypot links. 10. Email Security: Improved UI for email security demos and recipient modal performance. +30 other amazing updates and developments! We’re hosting an event at the upcoming Blackhat in Las Vegas in August; we hope to see you there. Keep an eye out for more updates, and feel free to reach out if you'd like to learn more or schedule a demo: https://lnkd.in/e92mkHyP. Have a great weekend, everyone! 🌟 #Features #Cybersecurity #Innovation #technology
Cadenza reposted this
We are thrilled to release FlashAttention-3 in partnership with Meta, NVIDIA Princeton University, and Colfax International The improvements from FlashAttention-3 will result in: • More efficient GPU Utilization - up to 75% from 35% • Better performance with lower precision while maintaining accuracy • Ability to use longer context in LLMs more efficiently https://lnkd.in/gx48DBuB
Cadenza reposted this
Here's what differentiates a great product security team from an OK one: How much they invest in the top right quadrant Any team can buy pentests. And all teams have to show up for IR But not all teams have the skills and tenacity to really invest in the proactive scaled quadrant But I think people will misinterpret what I'm saying here, so let me clarify: The top right quadrant is not just "security engineering" and scale is not just automation Scale is a superset of automation is a superset of engineering But there's lots of ways to achieve scale that isn't automation or security engineering Examples: - Great developer education - Building a culture of security across the company - Providing high-quality security documentation - Finding and providing pre-built secure defaults At Remy, we're on a mission to move the threat modeling and secure design review up from manual to scaled, so more prodsec teams have access to simple but highly effective proactive scaled programs
Technology has undoubtedly changed the legal industry over the past few decades. However, the impact of Artificial Intelligence is about to transform the more than 200-year old profession in ways never before thought possible. The Past: Working in the paper-based 1960s and 70s legal industry consisted of looking up the United States Code (USC) and relevant case summaries by hand in vast tombs that resembled a medieval codex. Needless to say, this was an expensive and time-consuming process. The Present: Fortunately for attorneys who passed the bar after the turn of the millennium, things got a little easier. The first digital systems, which were direct continuations of the paper-based ones emerged. This cut down the time it takes to conduct thorough legal research from a few days to no more than a few hours, but one lingering problem remained. The search function was and still is largely based on a set of esoteric identifiers that are less than ideal for ease of use and efficiency. The Future: Artificial Intelligence is changing this by enabling just about anyone to get the answer to any legal question they have at lightning speed, simply by typing out a search query in plain English. No endless scrolling through hundreds of categories or inputting identifiers that look like they could be part of a long-lost ancient language. Just one or two sentences and you receive a high-quality summary back. Some of these tools are already in use today, while others are just taking shape, like Marveri, whom we have funded, and who is creating an AI-driven platform that provides unique tools for corporate, banking, and finance lawyers. Learn more about how technology has changed the legal industry in our recent Cadenza Chronicle: https://lnkd.in/gxg-k5ys
Cadenza reposted this
Adoption is the hardest thing about new threat modeling programs. We generally see four modes: Release gates, consultations, security champions and full self-guided. Release gates. Engineering teams must go thru security to release. Strengths: Adoption/coverage guarantees. Weaknesses: Enormous overhead killing engineering output. Bottlenecks your whole business on security team headcount. Consultations. Engineering teams are encouraged to ask the security team for help when they are building something risky. Strengths: You get deep security expertise for the treat models. Weaknesses: No coverage guarantees, engineering teams are not risk triage experts nor incentivized to bring security everything they should. You are hamstrung by product security team headcount. Security champions. One engineer per team takes on a second role as security lead for their team and help with (particulalry) threat modeling. The key idea here I think is giving both incentive and education. Strengths: Embedded expertise and sense of responsability across the business. Weaknesses: The “two bosses” problem. Enormous initial rollout cost for finding and educating security champions around the business (and maintenance!). Fully self-guided. You ask everyone to threat model what they build on their own. Strengths: Low cost and effort to start. Weaknesses: If everyone is responsible, no one is responsible. Risk triage and threat modeling is not going to be everyone’s forte. Easy to do decently, but near-impossible to make great. I think all of these modes fail in their own way, but fully self-guided seems to work the worst. When a security champion program is executed perfectly, it really works, but that’s very rare. And all except release gates doesn’t actually really solve adoption, and the realease gates tradeoff is not one I would take without thinking twice. What I want (and what we’re building at Remy): - Engineering teams don’t need to proactively make risk triage decisions, but coverage should still be guaranteed. - The amount of secure design thinking isn’t throttled by security team headcount. - There is no security champion education or maintenance. Have you been able to get high adoption (say 80%+ of engineering plans meeting your bar for being “risky” get a proper review) with one of these modes?
The Cloud may be on the verge of its biggest innovation yet - Artificial Intelligence. But are the two applications compatible or a dangerous combination? Simply put, The Cloud and AI are a virtual power couple. On one hand, The Cloud is helping AI proliferate at the enterprise level by enabling all kinds of organizations from Fortune 500 companies to small non-profits to have direct access to it. In contrast, on the other, AI has the clear capability of enhancing the cloud in numerous ways. One example of The Cloud and AI combo is our portfolio company Together AI. Its platform is an early example of cloud-based generative AI in action, helping organizations implement AI tools quickly and cost-effectively in various cloud environments. But this is just the beginning of The Cloud and AI combo becoming a permanent fixture in enterprise environments worldwide. Learn how it's becoming the norm, rather than the exception: https://lnkd.in/g6edDN8h
Cybersecurity is now a $2 trillion industry. This is because according to the FBI's Internet Crime Complaint Center (IC3), annual losses due to cybercrime now approximate $12.5 billion. The answer lies in the very thing cyberattacks seek to exploit - data, and what is better at analyzing and learning from vast amounts of data than Artificial Intelligence? From having the capacity to identify statistical anomalies before they become attacks in real time to automating responses to incidents, AI is a threat to the shadowy multi-billion-dollar business of cyberattacks. Here is one example of how AI is being used to deal with online threats from one of our portfolio companies: Jericho Security Imagine this, it's the middle of the day and you're sitting at your desk when you receive an innocent-looking email from one of your vendors. It offers you a discount on products you are already purchasing from them and a link to buy them at a reduced rate. Thinking you're getting a deal, you excitedly click the link and your screen goes completely blank. You have just been hit with an automated phishing attack. Jericho helps to prevent such a scenario from ever happening by empowering organizations to defend themselves against the latest cyberattack techniques with its AI-powered cyber security training for employees. Aimed at mid to large-size businesses with dedicated IT staff already in place, Jericho's training platform provides hyper-realistic, personalized attack simulations with its phishing simulation platform. Given that generative AI's ability to mimic human communication and create 'deepfakes' is growing by the day, Jericho Security stands as a first line of defense against such emerging threats. To learn more about how AI is changing cybersecurity, check out Cadenza Chronicle's latest post: https://lnkd.in/gaySqmXv
Cadenza reposted this
WAGMI as we expand globally! Our recent acquisition of BitOasis is definitely cause for celebration. This strategic step, following CoinDCX's initial investment last year, solidifies our presence in the MENA region. The article also delves into BitOasis's remarkable growth story from its beginnings. Read the full article - https://lnkd.in/dqhMUcQc #CoinDCX #LifeAtCoinDCX #Bitoasis #Acquisition
Cadenza reposted this
The difference between owning BTC in an ETF vs holding your own keys may seem small if you haven't thought about it much - but in the situations where it really counts, you want to be holding your own keys.