Kurt Smith

In 2022, there was over 4,000 data brokers in the market. Today only about ~500 are registered in California's public registry (the largest of the 4). A old but great write up from Tom Kemp clocks the number of data brokers at around 4,000 - but the new legal definitions of Data Brokers according to Cali law is WAY broader! Any company selling, renting, releasing, disclosing, disseminating, making available, transferring, or otherwise communicating personal information of an individual they don't have a direct relationship with to a third party for monetary or other valuable consideration (including advertising) is a data broker now! Maybe time for an updated article, Tom? https://lnkd.in/dcqVJBRY

9

3 Comments

Today’s "From the Archives" Talking Logistics post was originally published in March 2021. It offers this warning: If any of your enterprise systems look like they were designed in the Windows 95 era, or make you take a Rube Goldberg approach to processing what should be a simple transaction, then you might be setting yourself up for a costly mistake. #supplychain #logistics #UI #UE #software https://lnkd.in/eYF8Zrf

6

1 Comment

You know those "free" solutions are more expensive, right?! I admit I've grown tired of being dismissed as a "software vendor". Any time I suggest a paid tool to a smaller revenue team, they dismiss me because "we don't have the budget" This is especially prevalent in the industrial and manufacturing space. 1 or 2 person marketing teams with an army of manufacturer reps. Everything is using manual spreadsheets and emails. You know what's worse? "Empathetic" consultants accept this fact and give them advice on how to do "more with less" They deliver workshops on how to do everything in spreadsheets: ❌ Jerry-rig a CRM ❌ Do Sales Forecasting ❌ Manage Support Tickets ❌ Manage Marketing Events You think you're saving money, but your manual processes cost you employee time And time is money You think you're saving money, but your manual processes are more error-prone Errors cost you time, and time is money So I guess you're not saving money, are you? "But Selim, what are some places we shouldn't skimp on tools?" 👉🏻 CMS: You're not saving money on WordPress, you're forcing marketers to hire developers, paying for hosting, and opening up yourself for security risks. Use the HubSpot CMS tier you're comfortable with instead 👉🏻 CRM: Spreadsheets are not the answer. Use the HubSpot CRM tier you're comfortable with instead 👉🏻 PM: Managing your projects and events should not live in one load-bearing spreadsheet that could crumble at any moment. Use ClickUp for tasks and time tracking. Use HubSpot for event attribution. Your free solutions are too expensive.

13

12 Comments

WSJ: California has become ground zero in the fight over artificial-intelligence regulation. AI startups and tech giants are rallying to kill a bill ascending through the state legislature that 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. The bill is widely viewed as crucial for how the technology will be regulated across the U.S., as California is home to many AI companies and often has an outsize effect on laws in other states. Proposals to regulate AI nationally have made little progress in Washington.

1

Brands - are your online reviews fake? How do you know? Can bad actors flood your products with AI generated reviews and get you in the FTCs crosshairs? Do you monitor your reviews? Lets start there - let’s help you get started at MetricsCart.com “The Federal Trade Commission on Wednesday voted unanimously to ban marketers from using fake reviews, such as those generated with AI technology, and other misleading practices to promote their products and services.” #ai #retail #reviewmonitoring #metricscart

5

Today's tech news highlights several exciting developments. Capcom is set to release the "Resident Evil 2" remake for Apple devices this December, expanding gaming access for iPhone and iPad users. GOG is preserving classic games for modern systems, with 92 titles now updated. Apple has launched Final Cut Pro 11, featuring new AI tools, while Bluesky's user base has surged to 15 million. EA's NHL 25 will include Professional Women's Hockey League teams, and IBM is enhancing quantum computing capabilities. Meanwhile, Spotify introduces a new payout model for video creators, and OpenAI plans to launch an AI agent early next year. Lastly, Sonos reports a revenue dip following challenges with its app rollout. This post provides a summary of key articles, aggregated, curated, and summarized by Mindlab’s AI Agent from trusted sources. For more in-depth insights, visit https://www.mindlab.news, or join Mindlab's LinkedIn Page: https://lnkd.in/gAM9bTYj The image below was created using MindLab's AI Agents, drawing inspiration from current technological developments reported in various online news sources. #GenerativeAI #AI #MindLab

3

Totally not tech related, but supremely concerning none the less - https://lnkd.in/g7DwqeJa "Due to the dwindling supply of quality machine shops and very poor quality aftermarket parts as well as a lack of people interested in learning the engine building and machine shop trades, I regret to inform our many followers and current as well as past customers that we will no longer be taking on any new engine build orders unless it is for a car we are restoring. Since Covid, we have had to do rework on multiple engine builds due to poorly manufactured parts that failed during break in or machine work that was below our standards due to all the old farts like me dying off with no younger workers interested in taking their place. For example, rod bearings are now made too thin resulting in 390 and 401 crank grinds needing to be ground .0085, .0185 or .0285 under standard rod journal size yet all but one machine shop in the entire Phoenix area refused to do anything other than the standard .010, .020 or .030 grinds. Even worse, when the one shop that will grind the cranks the way we tell them we need them loses their crank grinder to retirement in another year or two, they do not plan to replace him. Machine work that used to have a turnaround of 2-3 weeks now takes a minimum of 2-4 months due to an acute lack of people interested in learning machine work and doing manual physical labor. In fact, one engine block was at a machine shop for a year and when we got it back hey did such a poor sleeve job in one cylinder that it was not even useable so it is now a 250 lb paper weight.

6

1 Comment

How I Built an MVP App in a Few Hours Using AI Tools. 💡 Last week, I discovered something incredible: the perfect AI stack for rapid prototyping. Here's how I used Claude AI, v0, Cursor AI, and Replit to build a working MVP faster than ever before! The Challenge/Goal: Build a simple AI ROI Calculator App with minimal time investment. The AI Stack: Claude AI: My AI architect and coding companion v0: The visual prototyping wizard Cursor AI: My intelligent code editor Replit: The seamless deployment platform Here's the workflow process: ✅ First, I had Claude AI design the system architecture. Instead of spending hours on documentation, I described my needs conversationally. Claude broke down the components, suggested the tech stack, and even outlined the tech flows and structure. Time saved: Countless hours! ✅ Next, I used v0 to rapidly prototype the UI. I described my vision in plain English, and v0 generated a clean, responsive interface. The magic? It produced a production-ready solution using Next.js, Radix UI, Tailwindcss, etc. Time saved: 1.5 hours. ✅ Enter Cursor AI. With the architecture and UI ready, I let Cursor help me write the backend code. Its real-time suggestions caught bugs before they happened, and its ability to understand context meant fewer context switches. Time saved: 1 hour. ✅ Finally, Replit tied everything together. One-click deployment, instant collaboration features, and no complex setup needed. Time saved: 45 minutes. 🎯 The Results: Working MVP completed in 4 hours (use this link https://airoi.v0code.com ) Clean, maintainable code base Production-ready deployment Zero technical debt from rushed development 📊 Key Insights: AI tools shine when used together - they complement each other's strengths The human role shifts from coding to orchestration and decision-making Rapid iteration becomes truly rapid when AI handles the heavy lifting The future of software development isn't about replacing software developers/engineers, it's about boosting and amplifying their unique human capabilities which includes critical thinking! Using this stack, you're not just coding faster; you're thinking and building at a higher level! What's your experience with AI development tools? Drop in the comments below👇 #AI #SoftwareDevelopment #MVP #Innovation #Programming #TechStack #StartUp #CTO #CAIO #v0 #ClaudAI #CursorAI #Replit

10

2 Comments

James Bligh makes a great point in his medium post today. (https://lnkd.in/ggQczEyn) The ABA might hate CDR as their members for the most part seem to have formed a view that it is a threat and something that should be thwarted at every turn, or at the very least is yet another impost and regulatory burden they just needed to tick off their ever growing list of compliance projects. What banks and others may be overlooking is that the vast majority of Australians now have access to their banking data through a well designed and regulated infrastructure. Consumers and businesses don’t need to use screen scrapers and bits of string to get a picture of their finances. Locking in your customers so they can’t compare or migrate is a poor substitute for a strategy of just providing a superior service. Some banks get this and they will use the CDR as a set of ‘rails’ to drive transformations in customer experiences, customer acquisition and new value creation. At Trideca we are already working with several institutions that get this. The ones dragging their heels or who are still thinking that they are ‘done’with CDR now they have met the compliance hurdle are going to be in for a rude shock.

18

Rundown : trump may name an 'ai czar' to help coordinate federal regulation and government use of artificial intelligence. the role could be combined with a 'crypto czar' role. no names have been put forth for consideration, but elon musk and vivek ramaswamy will have input into who gets the ai czar role. Key Points: - trump's potential ai czar would help coordinate federal regulation of artificial intelligence - the role could be combined with a 'crypto czar' role to cover both areas - no names have been put forth for consideration, but elon musk and vivek ramaswamy will have input into who gets the ai czar role Why it matters : trump's potential ai czar appointment could shape the future of artificial intelligence policy in the us 🚀 Discover more insights by following @contentbuffer! 🌐 Stay ahead with the latest in AI, essential tips, and key developments! Join our growing community by subscribing to ContentBuffer’s free newsletter. Subscribe here: https://lnkd.in/eHhzDm5C #ai #trump #whitehouse #techstack #codinghumor #coding #cybercrime #ruby #techcareers #devhumor #devcommunity #cyberdefense #computerscience #mixedreality #softwaredesign #devlife #techtraining #programming #digitalnomad #naturallanguageprocessing #developerlife #iosdeveloper #AI

3

Search Wars -- Over the years we've witnessed the browser wars, the mobile wars and a broad assortment of CE format wars going back to the days when Cro Magnons fought over which club was better to kill a saber toothed tiger with... Enter the Search Wars 2 - with OpenAI diddling Google's "how you find stuff?" format and Google counterattacking... Questions abound, including how does OpenAI make money aside from being Microsoft's White Walkers? one question that leaps to mind is how long search will break for, catching web users in the crossfire. This isn't Alta Vista v. Google, where browsers were no more important than handkerchiefs - good search is as essential as clean water to most industries. https://lnkd.in/gvyy_-7g

21

9 Comments

"A later part of the plan to dominate the web, discussed at Microsoft but never executed, was to kill Google. .... After internal debate, Microsoft declined to kill Google not due to fear of antitrust action, but due to fear of bad PR from the ensuing antitrust action." Anti-trust action against big tech has decades-long effects on the technology industry. That wasn't the point of the https://lnkd.in/grJe4Wsg article by Dan Luu (excellent read BTW) but that's what stood out to me in the end. And a key criticism that is highlighted in the article is the fact that antitrust actions against big tech don't get the coverage that they deserve in the media, not compared to how it was when Microsoft was being investigated. "It might be hard to imagine this if you weren't around at the time, but the DoJ vs. Microsoft case was regular front-page news in a way that we haven't seen since." This is especially relevant today with the antitrust actions against Google in the US (https://lnkd.in/gcQnxgJz) and the regulatory pushback against Meta in India (https://lnkd.in/gaY_n4bC). P.S. I strongly believe that technologists must build a basic understanding of public policy. If you are interested in this intersection of technology and policy, you should check out the Graduate Certificate course on Public Policy @ The Takshashila Institution or the fellowship by Citizens for Public Leadership.

19

Who is reading all those meeting transcripts? I feel like we need an AI assistant to summarize all of the meeting transcripts for the day or week. I know this is coming. Don't get me wrong, AI transcriptions help with capturing decisions, action items and meeting summaries but reading the transcript is not likely. It will be interesting to see where meeting go in the future. All AI assistants? Southern California SIM (SCSIM) Society for Information Management #AI #CIO #toomanymaatings

16

11 Comments

I'm going to start including in my contracts with distributors that whiteboards are *mandatory* during my on-site visits. The bigger, the better. And please, make sure the markers work. 😂 I've been working on a framework for a custom intelligence platform for a client, involving internal requests, external customer requests, quotes/bids to orders, and analytics. Before you ask, YES, there's a GenAI LLM integration and RAG component. (It's actually going to be cool, we'll have a multi-prompt queue depending on the input - it's bordering on agentic - more on that later) It's been great discussing this solution with my client - but there were some outstanding questions on how this would all come together—mostly because this type of platform didn't exist previously. So it was difficult to demo how all the intended functionality would work in production. I couldn't show my client the final product end-to-end. Because it doesn't yet exist. And actually, that's typical of most of the work that I do. MOST of my engagements with clients involve software applications that don't yet exist in their final form. I have examples, boilerplates, mockups. But the final product that my client needs? Again, it doesn't exist—and if it does, I'm referring you to one of the increasing purpose-built-for-distributors software solutions out there, and the awesome people building those. The reason I'm talking with any distributor about a solution in the first place is because it doesn't exist off-the-shelf, in the way a distributor needs. So, we're designing it and building it - together. 🤝 It's not easy to build something completely custom: - I've been doing this so long that I can connect the dots, interpolate, and extrapolate. - Distributors are used to buying off-the-shelf solutions that they can "feature-shop-and-compare." I find that having a whiteboard, designing the flowchart, writing the requirements, and then connecting the dots together is an incredible way to help us all get on the same page. The last few days I've been on the go, so I didn't have a big space for me to put everything together. Thanks to a change of plans today, I found this whiteboard in a conference room downtown and got to work - and the solution started coming together. The same thing happens when we're on-site, designing a solution—a whiteboard plays a big role in that. And though I haven't done a causation study, it seems that the bigger the whiteboard we have, the better the end solution. 💯 This is why I'm going to start including a mandatory whiteboard clause in our contracts - so we can create better work, together. Oh, and if the markers don't work, it's obligatory that the person throwing it in the trashcan from across the room says "Kobe" 🙏 P.s. For a pro tip, you can actually take a picture of the whiteboard, and interact with it through ChatGPT - try it.

41

21 Comments

Generative AI is a 5 year-old goldfish. Why? Let’s think about it for a moment. “They say that goldfish have no memory…” - Ani DiFranco, Little Plastic Castles Okay, no memory (ignore the facts on this for a moment). GPT-4o and similar models, from the perspective of user interaction, don’t remember your conversations with them. Every time you start a new chat with an LLM, you have to start over and tell it who you are, your context, etc. Tools like ChatGPT with their memory feature layer memory over the top, but it’s not intrinsic to the model. Why is it a 5 year old? If you’ve ever raised kids, you know that some 5 year olds have the ability to follow multi-step instructions like “go upstairs, brush your teeth and get into bed so I can tell you a story.” Now, sometimes they only do part of what you ask, and other times they don’t do it at all… sound familiar? So to work with your 5 year old goldfish, you need to have a conversation something like this. “You’re a goldfish and I am your parent which means I can tell you what to do and you’ll do it. It’s time for bed and you like bedtime because I tell you stories which you really love. In order for me to tell you a story you need to go upstairs by putting your feet one after the other on the stairs (etc), brush your teeth (add instructions on that), and then get into your pajamas and then in the bed If you do all of that, step by step, I will come tell you a story.” And that, my friends, is how you can teach prompt engineering to non-techies…

25

11 Comments