Alex Pachikov

🚀 Nvidia’s $250 Jetson announcement is exciting, but the real story is the explosion of 12+ TOPS-capable SoCs priced under $200 that we’ve seen over the last 6 months. We’re about to see consumer products with MSRPs as low as $600 offering real, practical AI features: 🚗 Advanced Automotive: Think aftermarket Advanced Driver Assistance Systems (ADAS) with features like real-time lane centering and collision avoidance —at a fraction of the cost of built-in systems found in luxury cars today. Saferide 👀 🤖 Entertainment Robots: A smart TV that doesn’t give "personalized" recommendations but engages in a conversation about what you want to watch. Not sure if you’re in the mood for comedy or action? Argue it out with your TV—it learns your preferences, finds hidden gems, navigates across all your various streaming services, etc. 📑 AI Tutors: An AI tutor that helps you learn a language, prep for exams, or master coding. It adapts to your style, offers real-time feedback, and simplifies tough concepts—personalized support without the cost or hassle of traditional tutors. 🔐 Smart Security Systems: Affordable home security cameras with on-device AI processing for facial recognition, anomaly detection, and real-time alerts. No reliance on cloud processing means better privacy and no recurring costs—finally putting advanced security within reach for every household. This is the moment where AI becomes a daily, accessible experience for millions. What AI-powered consumer product would you love to see? 👇 #AI #ConsumerElectronics #Innovation #EdgeComputing #Nvidia #SoC #MediaTek #Rockchip #Qualcomm

25

2 Comments

Open access. Lots of talk about it, but what is it, how to do it? There is no getting around the need for ISPs to deploy some infrastructure; otherwise, they are just resellers of someone else’s service. But by sharing middle mile Infrastructure they can focus their Capex on last mile investments rather than duplicating what can be shared. The question then is, how to share it? In theory, Kenya's National Broadband Network is Open Access, but from what I've heard it's not a good solution. Leasing fibers on middle mile fiber would allow each provider to light it as fast as they would like. All of the civil works and fiber would be shared; the electronics is small portion of the total. In our transportation system, we share the roads, built by the public sector, and let multiple providers, that is private cars, taxis, buses, and trucks, all share the same infrastructure. That's a rough analogy, but maybe useful.

40

28 Comments

2 days ago California State Senate passed SB 1047, which establishes safety standards for AI developers of frontier models. A lot of the requirements directly demand guardrails that ZenGuard AI is developing. SB 1047 has far-reaching implications as it will inform other regulations in other states and countries. https://lnkd.in/dNSfJqun * All developers are strongly encouraged to follow safe, secure, and ethical AI development. The measures are (among others): pre-deployment safety testing, red-teaming, cybersecurity, safeguards to prevent the misuse of dangerous capabilities, and post-deployment monitoring. The goal is to set a community standard for safe, secure, and ethical AI development, that even smaller developers can adopt. * SB 1047 heavily targets its regulatory focus on "frontier" AI models requiring computing power greater than 10^26 floating-point operations that cost over $100 million to train. * The bill introduced CalCompute - a state-backed cloud computing cluster to democratize access to GPUs for smaller developers. In conjunction with the standards and guidelines in the bill, there is a clear motivation to adopt safety and security at the earliest stages of AI development for all developers. * This sets legal framework for attorneys to take legal action in the event the developer of a powerful AI harm or threat to the public safety.

19

4 Comments

Do you live in Seattle? Are you interested in Seattle's crime reports? Seattle makes its reported crime data publicly available on the city's website. The only problem is the dataset from 2008 to the present is 1.1M rows. That's a lot of data for the average person to manage and analyze. They can't simply open it in Excel because it's too big. To help, we've imported the entire dataset into a Row Zero spreadsheet and analyzed a number of different trends. If you want to take a look for yourself, you can open the data in Row Zero (for free) and drill down into your neighborhood to look at specific types of crimes. https://lnkd.in/gNcTqAwW Below are some noteworthy examples. Keep in mind that the Seattle city population decreased slightly during the pandemic and has since recovered, up about 5% from pre-pandemic levels. So small increases in crime would be expected due to population growth. - Total reported Seattle crime has decreased compared to pre-pandemic with notable drops in 2023 and 2024, even as the population has increased. Reported cases in 2022: 75.044k, 2023: 70.179k, 2024: 62.152k YTD. - Homicides are up significantly over the last 5 years. The last 5 years are the 5 worst years in the data, with reported homicides up more than 75% in the last 5 years. - Motor Vehicle Theft has nearly doubled since pre-pandemic with 2023 seeing a sharp spike in Seattle car theft. - Drug Offenses are up significantly in 2024, possibly reflecting stricter enforcement. However, they are still well below pre-pandemic numbers. - Despite national media reports of increased shoplifting, reported shoplifting crimes continue to be well below pre-pandemic levels, down more than 20% vs pre-pandemic. - Top 5 Worst Seattle Neighborhoods for Car Break-ins QUEEN ANNE (50% more car break-ins than any other neighborhood) DOWNTOWN COMMERCIAL UNIVERSITY SLU/CASCADE ROOSEVELT/RAVENNA Row Zero #Excel #spreadsheets #dataanalysis #seattle

32

2 Comments

California Kills AI 💀 This is a provocative headline no one wants to come true, but don't tell that to California's State Legislature. SB 1047 has been passed by California’s Senate and two of the state’s Assembly committees with little opposition. The Bill requires that developers of large AI models conduct safety tests to reduce the risks of “catastrophic harm” from their technology, which it defines as cyberattacks that cause at least $500 million in damage or mass casualties. The developers also must ensure their AI can be shut down by a human if it starts behaving dangerously. The Bill applies to AI models that meet a certain computing-power threshold and cost more than $100 million to train—the estimated cost of training OpenAI’s GPT–4. This criteria would almost certainly snag Microsoft Meta Google Oracle IBM Tesla and other IT giants who are developing GPT-4 equivalent or more powerful #llm #genai and #ai models. Any company doing business in California is covered, regardless of where it is based. AI #startups and tech giants are rallying to kill the Bill, which they say would impose impossibly vague constraints in the name of safety. Though some in the industry have called for government regulation, they say it should be done at the federal level with more specificity.  AI regulation has not made much headway in Washington. Key Thoughts: > AI deserves regulation, but the regulations probably should be at the federal level and they shouldn't happen until actual legislation is passed calling for regulation. > Given the recent Chevron ruling by the Supreme Court, which dramatically curtails the authority of Federal Agencies to act without legislation to expand federal oversight by unelected officials it is likely we need legislation for any regulations to be enforceable. > Businesses are right to push back on State Governments who are passing legislation like California SB 1047 because the states lack the competence to regulate industry for the entire country, especially in areas as complex and consequential as AI. > AI has the power to transform. We need to give #innovation some breathing room. It is true that giant steps have been taken in AI in a short period of time, but there hasn't been widespread harm, mostly fear. > We need to promote AI competition and innovation. The technology increases in power and capability every few months, while also becoming less expensive and accessible. Curtailing technology this powerful could significantly impact economic progress, especially in aging societies. #ethics #ceo #cfo #cio #leadership Terzo Credit: Wall Street Journal reporting by Preetika Rana

11

Today we share Nominal with the world. We are on a mission to revolutionize the way mission-critical hardware systems are tested, validated, and monitored. We’ve been hard at work for the past 18 months building our core product offering, working with early customers across commercial and DoD applications, and assembling a world-class team. With $27.5M in funding, we are hitting the accelerator and scaling. Thank you to Lizette Chapman at Bloomberg for sharing our story here! https://lnkd.in/eJQVwf9P The Nominal platform gives engineers and operators a first-class way to organize, visualize, transform, validate, and monitor data—all in one application. Flexible. Scalable. Collaborative. We are supporting the world’s most advanced hardware companies with best-in-class testing and analysis workflows. Our customers build and operate satellites, rockets, drones, autonomous aircraft, advanced nuclear reactors, and more. Our mission is to support their mission. Our early traction as a team is the result of years of patient observations, experience both building and operating software in diverse and demanding environments, and earned insights from working with some the best companies in the business. I’m thankful to be standing on the shoulders of giants. If 18 months is any indication, it’s proving to be the fulfillment of a lifetime. And we are just getting started. For me, Nominal is the conversion of a massive amount of potential to kinetic energy. I’ve spent a decade operating at the intersection of mission, technology, product, and company-building. My time in uniform as a US Navy submarine officer and later as a startup operator (Anduril Industries, Applied Intuition, Saildrone Inc) in critical technology areas has personally cemented a powerful truth: engineers and operators who are testing, validating, and monitoring high-stakes hardware systems deserve the world’s best software. I’m proud to be on this journey with Bryce Strauss and Jason Hoch—I can’t imagine two better people to motivate and inspire me every single day. Finally, we are lucky to be backed by some of the best investors in the world—thank you to Lux Capital, General Catalyst, Founders Fund, XYZ Venture Capital, Haystack, Human Capital, BoxGroup, and Overmatch. Your support from the earliest moments has propelled Nominal into a bright future! All systems Nominal. 🚀 Bryce Strauss, Jason Hoch, Paul Kwan, Alexa Liautaud, Bilal Zuberi, Josh Wolfe, Delian Asparouhov, Trae Stephens, Ross Fubini, Semil Shah, Aashay Sanghvi, Matthew Colford, Baris Akis, Adina Davis, Claire Smilow, Evan Loomis

908

166 Comments

It's holidays but Voice AI industry is still booming! 🔥🔥🔥 - Voice AI Agents market to reach $47.5B by 2034 - Hume AI announced a new AI voice model - PLAUD.AI brings AI note-taking assistant for all meetings and interactions - Reverie Language Technologies a voice API platform supporting 12 Indian languages - Engineered Arts Ltd secures $10M to scale up humanoid robots - Hugging Face released Moonshine Web: A browser-based privacy-focused STT - and much more All details here 👇 https://lnkd.in/eiK6ZUQj

66

4 Comments

Ethernet always wins. It has already adopted most of the features of Infiniband. Now, Meta reports Ethernet is more reliable.

14

What Happened to Orbital Insight? 🛰️📉 Orbital Insight, once positioned as a pioneer in geospatial analytics with backing from major investors like GV (Google Ventures) and IQT (In-Q-Tel), set out to revolutionize data analysis with its TerraScope platform. The company raised over $125 million to apply satellite imagery and AI for everything from monitoring global supply chains to military surveillance. 🚀💼 However, by 2020, the company’s commercial revenues began to fall, and despite securing emergency loans and cutting staff, Orbital Insight struggled to regain its financial footing. Failed merger talks and mounting financial pressures left the company scrambling for stability. Even though Orbital secured key defense contracts with Indonesia, Taiwan, and Israel, they’ve had difficulty finding a sustainable path forward. 📉 The Lesson? Orbital Insight’s journey reveals how even the most cutting-edge tech innovations need more than just big ideas—they need financial resilience. While partnerships and collaborations with major tech or defense contractors could have provided a safety net, it’s clear that their challenges went deeper. Falling commercial demand, market shifts, and internal operational challenges all compounded the company’s struggles. While partnerships might have helped, Orbital’s story reminds us that a solid financial foundation, adaptability to market conditions, and strategic risk management are crucial for any high-tech venture looking to scale and sustain success. We're exploring some really interesting companies, hope you all are enjoying learning as much as I am :)

1

Harry Sohie is on a 🔥 streak this month, building big features in WholeCell.io. We just launched THREE updates to Scan Reports. My favorite of which is automated scan actions. If you're buying phones in #telecom auctions, WholeCell now has a fantastic workflow for receiving your inbound purchase orders. Scanning an IMEI or SN can automatically update an item's status and location, recording who scanned it and when. This saves so much time for operations teams, and should be an #R2 auditor's dream for reliable, standardized historical data.

15

2 Comments

California Law SB 1120 has gone into effect, and prohibits health insurance providers from denying claims solely based on an algorithmic decision. This means that "yes" decisions can be automated, but "no" decisions must be routed to a human expert reviewer for a final decision. The law is intended to disallow mass automated rejections, as has happened before, and matches how most health insurers already operate. https://lnkd.in/gKDx_XVa #healthcare #ai #regulation

42

4 Comments

💻🤝 Cephable is a personal, private, and accessible AI platform that enables users to control their digital devices through adaptive voice commands, head movements, virtual buttons, and AI task automation 🤝 💻 It's W12 / Q4 / 2024 & this week in our MGMT Boston startup series we’ve got Cephable, a Seed stage startup building a platform designed to enhance accessibility and inclusivity. Cephable was founded in 2023 by Alexander Dunn, an applied ML engineer who was formerly the CPO at Voicify. A tinkerer by trade, he saw that his younger brother, who has a disability, was unable to keep up with friends while gaming, and falling behind with the technology at school. Globally there are more than 1.3B people (15-20% of the global population) with disabilities who have an annual discretionary spending power of $13T according to the World Economic Forum. This includes cognitive, mobility, hearing, vision & other restrictions that prevent millions of people from enjoying games & digital experiences. 💵 Can you imagine not being able to access an ATM to take out some cash?  🐢 Or fire a tortoise shell at Wario on the Rainbow Road racetrack?  ✉️ Or how about just logging onto the computer to send an email? Cephable is enabling a new digital customer via their software platform which launched publicly in August 2023 to solve the “disability tax,” an additional cost that many people with disabilities have to pay for adaptive tech. Today they’ve built innovative, responsive algorithms that are adaptive to visual & audio commands across languages and commands. Cephable recently announced partnerships with Microsoft & Qualcomm to build “out of the box” PC experiences which will include Cephable controls integrated into OEM (original equipment manufacturing) devices like Lenovo, HP & Dell computers & tablets. These large scale enterprise partnerships will take the startup from 10s of thousands of users to 10s of millions in the coming quarters. This fast-growing Boston-based startup grew their end user base from early adoption to its 2024 year-end user base by more than 10,000x. They’re expecting to grow another 10,000x in 2025, turbocharged by their newly established enterprise relationships. They are closing in on seven figures of annual revenue in the 1H of 2025, with their average Enterprise pipeline deal size now approaching seven figures too! Next year they will plan to bring on more engineers, marketing resources, and grow their customer success & sales operations teams. Cephable is backed locally by Hyperplane, Founder Collective, and other notable VCs. Operators to Know: -Jason Fields, COO -Jonathan Campbell, Director of Engineering -Cordellia Yokum, Senior Software Development Engineer -Shristi Chitlangia, Software Engineer -Julia Franklin, Chief Learning Officer -Alexa Orban, Director of Marketing Communications If you're interested in learning more, including the key roles Cephable is hiring for, the full post is linked below

87

1 Comment

'OpenAI co-founder Ilya Sutskever's new safety-focused AI startup Safe Superintelligence (SSI) raises $1 billion' "Safe Superintelligence (SSI), newly co-founded by OpenAI's former chief scientist Ilya Sutskever, has raised $1 billion in cash to help develop safe artificial intelligence systems that far surpass human capabilities. The company declined to share its valuation but sources close to the matter said it was valued at $5 billion." "Investors included top venture capital firms Andreessen Horowitz, Sequoia Capital, DST Global and SV Angel. NFDG, an investment partnership run by Nat Friedman and SSI's Chief Executive Daniel Gross, also participated." ""It's important for us to be surrounded by investors who understand, respect and support our mission, which is to make a straight shot to safe superintelligence and in particular to spend a couple of years doing R&D on our product before bringing it to market," Gross said in an interview." "AI safety, which refers to preventing AI from causing harm, is a hot topic amid fears that rogue AI could act against the interests of humanity or even cause human extinction." "Sutskever, 37, is one of the most influential technologists in AI. He co-founded SSI in June with Daniel Gross, who previously led AI initiatives at Apple, and Daniel Levy, a former OpenAI researcher. Sutskever is chief scientist and Levy is principal scientist, while Gross is responsible for computing power and fundraising." #SSI #SafeSuperintelligence #AI #artificialintelligence #machinelearning

31

3 Comments

The latest edition of The Driverless Digest has just been published! Highlights from this week: 🤖 Waymo now at $45 Billion valuation but how do they stack up to Uber at a public market cap of $150 billion? 💵🔥Zoox to launch in San Francisco and Las Vegas. 🤖 Waymo working on scheduled rides feature which is great for customers but read why I think it’s only going to hurt their utilization. Check it out here - https://lnkd.in/gNywsSmW

18

Uncommon opinion: call centers should worry less about detecting deepfakes calls. "Deepfake call detection" really just means "detecting synthetic audio." But there are lots of reasons to synthesize audio! It could be that you literally cannot speak, and are relying on voice synthesis to be able to engage with the world. It could be that you used sophisticated noise cancellation to make yourself more intelligible. It could be that your biological voice induces a sense of dysphoria and you're using a voice converter to obtain another that's a better match for your identity. The danger isn't really synthetic audio - it's lies, manipulation, fraud, etc. But searching for signs that the audio was digitally processed won't tell you whether it's being used maliciously or not. Why not just build a solution to the actual problem? Instead of deepfake detection, let's focus on detecting malice, manipulation, and deceit. As a bonus, this protects us from bad human actors too.

35

2 Comments