Brian K. Buntz’s Post

Woke up to some fascinating news this morning: Anthropic has rolled out their latest LLM models (Claude 3) in three different flavors, all of which appear to be "smarter" and faster than ChatGPT, and 2 of which appear to be cheaper (reference: https://lnkd.in/gPSJWjKH). This has some big implications for Locusive. If these models are truly as good as, or better than, ChatGPT, particularly in their textual reasoning abilities, then the pricing schema for Locusive might become dramatically simpler and more straightforward. Before, I was really straining to try to figure out how we might price our system, which currently averages anywhere from 20K-30K tokens per request, in a way that allows for predictability and doesn't cost an arm and a leg. If we can use Anthropic's latest systems, we might be able to move to a much simpler and more standard subscription model, potentially one with a generous free tier. A lot remains to be seen, particularly whether this model is smart enough to act as the brains behind an autonomous agent, but I'm excited. --- One other thing that strikes me, today, is my reaction to this post, which also mentions the ability to have better agents and more context, with my reaction to OpenAI's similar post from a few months ago. Back when OpenAI announced stronger integrations with enterprise apps and an agent model, I was freaked out. What they were announcing, at least at surface level, seemed like it would obsolesce what I was doing. Fast forward a few months later, and we've had the chance to see their Assistants API and their enterprise data integrations. While impressive, it goes to show that there's still a business to be had for folks focusing on vertical use cases, and I don't think that's going to change. I believe that, while they may have made it easier for folks to try LLMs with their own data, there's a lot of work to be done in making application-specific use cases seamless. My reaction today is one of excitement and hope at the possibilities, rather than concern, fear, or worry. I think it's a good time to be a software builder.

I've come to realize just how deeply sophisticated bots have infiltrated the internet. They're not just lurking in the shadows; they're actively shaping conversations, setting up elaborate ruses where they simulate entire debates. It's like watching a play where all the actors are bots, each presenting different views, making you think you're seeing a genuine dispute. You might feel like you're forming your own opinions, but the whole scenario has been crafted to guide you to a predetermined conclusion. It's eerie, really. For a while, I thought the notion of a digital apocalypse was just dramatic talk, but my views shifted dramatically today. I stumbled upon something on Reddit that blew my mind. I knew bots were common there, but I didn't realize the extent of it until I saw a screenshot. This screenshot wasn’t just any repost—it showed the same discussion thread recreated ten months apart, complete with the original posts, user comments, and even replies. The exact same interactions, just different timestamps. And the kicker? No advanced large language models were needed to pull this off, which means this trickery could be just the tip of the iceberg. It makes you wonder what else could be automated without us knowing. Here’s a free startup idea for you—no need to thank me, just promise to think critically about the 'human' interactions online! Credit Notes: Sergei Smirnov and Telegram Group: ai_newz

GenAI hot take... TL;DR - was a fan, then a detractor, now seeing the magic at last Like many, I've experienced the ups and downs inherent in a new technology paradigm like generative AI. It started with consistent 'jaw drop' moments when ChatGPT hit the stage (like writing personalized stories for my kids), but that enthusiasm gradually devolved into a sense of worry. The profound potential of the tech was always there, but the personal frustration (evolution after that initial big bang was slow) and professional doubt (too many companies stuck in pilot purgatory and not seeing REAL value) kept building. But through a nice bit of serendipity, I've had no less than a handful of 'it's finally here' moments in the last week alone. Just this morning I was messing around with creation of a new website. Nothing fancy, just a 'Coming Soon' type page for a new venture (stay tuned on that!). Where GenAI-fueled site builders in the past would create more work than they alleviated, today with a few prompts I got a simple site spun up in under 5 mins (kudos on the progress, WordPress). Perfect? Not even close. But a strong signal we are starting to crest the hilltop and create real value from this massive wave of investment and hype? ABSOLUTELY. #ai #genai

🚀 300M Weekly Users. 1B Daily Messages. And AGI by 2025? Sam Altman just shared some jaw-dropping insights at the NYT Dealbook Summit: 300 million weekly active users on ChatGPT, exchanging over 1 billion messages daily—a testament to how rapidly AI is becoming a part of everyday life. His prediction? AGI glimpses by 2025. While the initial effects might feel subtle, Altman warns that the transition to superintelligence could be “more intense than people think.” Other key takeaways: Microsoft partnership: Acknowledged "some tension," but they remain aligned on priorities. On Elon Musk: Described the OpenAI legal issues as "tremendously sad" while calling Musk a childhood "mega hero." He also expressed confidence that Elon’s growing influence wouldn’t be wielded to harm competitors, calling that approach “profoundly un-American.” Sam’s interviews are always fascinating and look into the future, and this one reminds us how quickly AI innovation is accelerating. What do you think? Is AGI by 2025 realistic—or overly ambitious? - - - Follow Abdul Rahim Roni for more updates 🔥 --- If you're tired of paywalls holding back your reading experience, try ProReader.io • If you like this and want more AI tools, tutorials, and news, join Cool Tech Gadgets - my newsletter that helps you get smarter about Microsoft AI Google and tech: 👉 https://lnkd.in/grn4UX-8.

We’re launching new AI features with Gemini in #Gmail! Tune into #GoogleIO on May 14 to get a sneak peek of what’s coming to your inbox →

What does the future of #IT look like? Not even ChatGPT can tell you for certain. But, get an overview of the top information technology trends for the coming year in our newest blog post:

What does the future of #IT look like? Not even ChatGPT can tell you for certain. But, get an overview of the top information technology trends for the coming year in our newest blog post:

Every company: “We’re AI-powered.” Everyone else: 🙄 Our team saw a lot of this kind of “AI-washing” at the AWS Reinforce Conference recently - with nearly every single sponsor proudly referring to themselves as AI-powered. But this designation means nothing unless the companies who claim it can keep these kinds of promises to their customers: 🔑We won’t let your data show up anywhere else. 🔑We won’t share your data with any third-party AI tools. 🔑When we change our underlying AI model, it won’t create unintended consequences for your product or business. 🔑We won’t introduce bias that goes against your company's values or goals. Of course, even if these companies make pinky promises to all of the above, there’s no way they can make these guarantees if “AI-powered” is just shorthand for “using a ChatGPT widget.” How could they…they can’t direct the actions of a third party like OpenAI! To truly protect customers’ data, a company’s AI usage should be proprietary, citation-heavy, and adherent to the latest certifications. That’s how we do it at TrustCloud, which means we’re comfortable saying yes to the whole list. So, yes, we are “AI-powered”...but we have the receipts to back it up! 🧾

🤖 What's the best AI tool these days? 🤖 I sheepishly asked Gridline's Head of Technology yesterday, Peter Bilali, over Farm Burger. He walked me through some of the options beyond ChatGPT. Today's The Wall Street Journal breaks things down even further across five bots for nine categories - e.g., summarization, health, wealth, and work writing. Article spoiler alert: ChatGPT is their #2 pick overall. See the table in first comment. Gridline's CEO, Logan Henderson, just used summarization for one of Gridline's soon-to-launch products, a custom fund focused on ______! Does your company want to set up a custom fund focused on any of the below? Contact me. Gridline can do it faster, cheaper, better, and with fewer operational headaches than other options. More info: https://lnkd.in/ebMsUukg 1️⃣ Managers YOUR firm likes 2️⃣ A certain sector of the economy or theme 3️⃣ Emerging managers 4️⃣ a specific DEI theme 5️⃣, 6️⃣ & 7️⃣ etc etc etc NOTE: We also have a GP feeder solution that helps IR/BD people pursue HNW/Wealth Management channels with their COO/CFO's blessings—instead of cursings! 🇺🇸 Happy Memorial Day Weekend, AND Never forget our heroes! 🇺🇸

On one hand, this doesn't change anything. Brands will still have to publish content and rely on reviews and (pay to play) listicles to feed the models and search engines. On the other hand content interfaces that are just wrappers around a relational DBs will adapt or fall to personalized chat interfaces (G2 crowd etc.) Google still has the upper hand here until publishers realize closed proprietary data is more valuable than partially open proprietary data. This will require a change from a content mindset to a service mindset. Also, authentic first party reviews from the source will weigh more than ever. Search will decline so brands will need to push out their reviews in discovery interfaces (think Monday.com but real customer stories). This is already prevalent in the DTC space. What a fun time with so much opportunity in a shifting space! Happy to help you think about your brand's next moves. 🤙 #seo #customerstories #contentmarketing #saas #growth #llm