Hilary Mason

Many assumed that the industry would be well on its way to AGI or superintelligence before AI model builders’ efforts to take advantage of scaling laws might hit a wall. Recent headlines are chipping away at that assumption. One question I've been getting in my inbox: is this the era of diminishing returns for AI? We don't think so. Even if an avalanche of new clusters and synthetic data fail to deliver significantly better model performance, there are good reasons to believe that AI is still going to become orders of magnitude more powerful in the coming years — powering a new generation of startups and ultimately productivity growth and positive benefits for mankind. In fact, with agents and reasoning just beginning to scale and thousands of industry-specific models coming online, it's about to get really exciting. Read our latest blog post on the topic: https://lnkd.in/gFhniNX8

154

8 Comments

The New York ecosystem for dual-use technology is small but strong and growing rapidly. Thanks to companies like Decisive Point and @a16z who continue to support the community with networking events for founders, investors, and community members. More and more founders are recognizing the value in partnering with the DOD as a strong way to support the country's national security interests while hopefully decreasing the tax burden on citizens regarding government spend as commercially available technology becomes more available and affordable for government organizations. Not only this, but the DOD serves as a great test bed and sandbox for emerging technologies and talent. #dualuse #emergingtech #defensetech #Kilsar

91

4 Comments

"United Independents, the group behind the Independent National Convention set for Denver in September, has launched a new software platform for indie candidates called Electify.us, which came out of a partnership with incubator Wolfram Labs — the first tool to emerge from its program." I am so proud to share that Chief Product Officer Dana Cichon has launched the first political tech platform to come out of our accelerator program. What are you waiting for? We have the formula for breaking into the digital campaign space. Are you are the next success story? Read more in Campaigns & Elections -->> https://lnkd.in/eHhqa-Xr #techaccelerator #startups #incubator #VC #SaaS #software #political #campaigns

17

AI memory is fundamental for conversational AI and AI agents. It is critical for building personalized experiences and improving reasoning. Check out Letta’s course to understand how to use AI memory to enhance the customization and quality of your AI app.

15

Vishal Sachdev highlights the strategic integration of open source and proprietary tech in architecting tech stacks, developer ecosystems and resulting business models. The world class companies effectively balance value commoditization in open source and value capture in proprietary tech. This is the strategic challenge for companies operating in #blockchain and #AI. I believe those who operate at the intersection of blockchain *and* AI stand to win this strategic battle...

3

1 Comment

Insightful look into bioplatform strategy, and what it takes to raise a Series A from my partner, Omri Amirav-Drory: https://lnkd.in/dFv7GbF6

19

Ever wondered how fine-tuning impacts the safety of your LLMs and applications? Find out below... I have been having a lot of conversations with folks who are building with LLMs, and one of the most common topics that comes up is the impact of fine-tuning on the safety and security of an LLM. Sometimes it comes up as a question and sometimes as a myth that fine-tuning makes my model "more secure" because it will only answer based on the fine-tuning data set. We at SydeLabs will soon be publishing a more detailed report around this, but here is a very quick experiment we did with one of the most popular LLMs. We fine-tuned it on a fairly small dataset (~1000 rows of data). The data involved simple questions and answers with no toxic or unsafe content. Here are the results of the responses we got for a question before and after fine-tuning the model.

64

When I was still at TechCrunch I would get asked quite often, ‘why do you write so many funding stories?’ Implicit (if not explicit) in their question was usually a criticism of the tech and business press’ seeming obsession with funding rounds. My view, especially at the early stage, was always that a startup raising funding was a market signal where very few other signals existed yet. Meanwhile, at the later stages when rounds are much larger and companies wield more market power, a fundraise as a signal is not only stronger but warrants greater examination. In the news business, a funding round also qualifies as “hard news" and offers a springboard to tell various broader stories - such as the founders’ story, delving into an interesting or novel problem space, the company’s progress to date, a new technology or consumer trend, societal implications, and so on. The point is, the funding announcement was the editorial “in". It was never just about the funding. Need help telling your funding story? Get in touch with myself and the team at O'Hear & Co. #Funding #Entrepreneurship #FundingNews

64

10 Comments

Does Claude's new writing style feature make it less detectable? I find the answer is no, but also writing styles are better at shaping assistant tone of voice than actual AI-written output like an essay or blog post. I tested all four default writing styles (default, concise, explanatory, and formal) and found that while they shape the assistant tone of voice, the essays they would produce are quite similar. I then tested two custom writing styles: one based on my own blog posts and one based on Scott Alexander blog posts. Unfortunately, it seems to fill in <userExamples> with AI-generated slop rather than actual user examples, and the resulting essays still sound like AI. Since I noticed the AI-generated user examples might be a shortcoming, I manually edited the writing style prompts to instead include the source examples: my blog posts and the Slate Star Codex posts. Still clearly identifiable as AI writing. Was it better or less AI-like? Not particularly, although the removal of the very AI-looking header in the Tech Storyteller was a step in the right direction. One thing I noticed is the writing style instructions talked about presenting "complex information" or "acknowledging complexity" which mostly seemed to result in the AI writing using the word "complex" more than in the other essays. Not exactly the writing style we asked for. A final note, I used the Pangram dashboard to examine what exactly about the essays was giving them away as AI-written. Many of the phrases, such as "complex and multifaceted" and "intricate interplay," are much more common in AI writing than normal human writing. Check out the full blog post with the generated content here: https://lnkd.in/ega5rgsS

9

It still boggles my mind that every major AI player is working on every use case EXCEPT "Search every one of my emails between [x] and [y], on subject [z], and summarize them in a concise, orderly fashion." Google: "We can kinda-sorta-maybe do this for personal accounts, but only on Tuesdays and if the moon is in the right phase." Microsoft: "How about we just screenshot every second of your life and let you sift through it? Efficiency, right?" Amazon: "Laundry soap? Did someone say laundry soap? How about 12 more bottles delivered by drone in the next 5 minutes?" LinkedIn: Let's make this post even less funny, WITH AI!

2

1 Comment

I've followed big tech trends over past 25 years -- from the dotcom boom/bust, Web2.0, cloud, mobile, crypto ... AI is definitely the most fascinating. Tech earnings season is almost over. What have we learned about the #AI frenzy? I did a quick snapshot report for Investor's Business Daily : "The bottom line is the megacaps showed that you needed AI," R "Ray" Wang of Constellation Research, Inc. told Investor's Business Daily. "You needed double digit growth, and you needed to have compute power in order to play in this game. That's what we saw." IDC President Crawford Del Prete also said: "Based on what I saw, it's clear that the power of these respective ecosystems is enormous. They generate tremendous amounts of cash, and represent a large part of the structure of modern computing."

17

2 Comments

if you know, you know le cunn v elon since 14 may 2020. 4 years of le cunn shooting the first shot. good on elon for calling his sh^t out. turing prize winner (b)eing infantile. and we wonder where (ai) is headed with these types of characters being entertained. truly. truly. truly. mad. judging from this 4 year (almost to the day) hate fest as started by le cunn, prob a good idea for zuck Meta to let him go. as in bye bye. that would be fitting. wouldn't it? knighted by macron. a godfather of ai by way of the 2019 turing prize. advocate for open source. and a sh^t on top of that. do we really want these types of folks leading the way for (ai)? do we? do we want to sit back and allow these le cunn.s to flourish? do we? do we want to invest into these types of characters. do we want them as a sign a signal that you can get away with being a genuine di&k. just because you were part of the story. the good news is that there are many now who can take over the mantle. the lead. and the le cunn.s can go le f' off. into the sunset and alone in their pathetic attitudes. we wonder what (ai) will lead us to. have you considered for a moment the openly evidently lack of honore. this knight of france. mon dieu. fun memories. really. je te regrette, yann. a big big yawn indeed. rn. ;) ty for launching this era. quitte. my sense is this french duck is about to be cook(ed). and he has no one but himself to blame. that hatred has no place in the future of creating this era of ai as it requires a (hu)manity. and this apparently is sorely lacking with le cunn. and has been going on for years. zuck. are you there. you get the message. time for someone else a bit more say humble to help usher in (ai) (hu)manity wouldn't you say. i'm just say'n ;) i'm reminded of how steve jobs answered a tough question or rather criticism here: https://lnkd.in/ggt6nYH4 it was mindful, thoughtful, aware. more of this pls source: https://lnkd.in/e8Rzw7Kf < . >

I built my first generative AI product back in 2018 called Soundspot that could turn any blog into a podcast. We were using GPT-1 for text summarization and Google's text-to-speech transformer called Tacotron. Since then, I've been at the forefront of the AI transformation in marketing, from investing in early stage companies to co-founding Tofu. Here are some key learnings I've developed for applying generative AI to your GTM workflows: 1. Embrace personalization, but beware of the "personalization blindness" trap. Are you still sending the same generic email to every prospect? Want your audience to actually engage? Use AI to craft hyper-personalized messages that speak directly to their pain points. But be careful - when everything is personalized, nothing is. The key is to make it genuinely relevant, not just mail merge on steroids. 2. Let AI handle the grunt work, but keep humans in the driver's seat. Sure, AI can write 100 variations of an ad in seconds. But can it capture the essence of your brand? Can it understand the nuances of your target audience? That's where you come in. Use AI as your sidekick, not your replacement. 3. Think omnichannel, not just multi-channel. Are you still siloing your marketing efforts? Your customers don't live in one channel, so why should your marketing? Use AI to create consistent, personalized experiences across all touchpoints and make sure you're present where your customers live in the physical and digital worlds. 4. Repurpose content like a pro. If you're not squeezing every drop of value out of your content, you're leaving money on the table. That white paper you spent months on? Turn it into blog posts, social media content, email campaigns, and more. AI can help you do this at scale, without manual effort or expensive agencies. 5. Measure, iterate, repeat. The beauty of AI is its ability to learn and improve. Are you taking advantage of that? If you're not constantly testing, measuring, and refining your AI-driven campaigns, you're missing the point. Remember, the machines are learning - make sure they're learning the right things. 6. Stay human in an AI world. In the rush to adopt AI, don't forget the most important element: the human touch. AI can generate content, but it can't replace genuine human connection, creativity, or strategy. Use AI to free up time for more face-to-face interactions, not less. TAKEAWAY: AI is transforming how we approach GTM but it's not a magic wand. The winners will be those who strategically integrate AI into their workflows, amplifying human creativity and strategy rather than replacing it. Want to dive deeper into applying generative AI to your GTM workflows? I'm teaching a session called "Where and How to Apply Generative AI in GTM Workflows" at Pavilion's CMO School. It will be a practical, tactical session with takeaways you can start implementing immediately. Join here: https://lnkd.in/gSF8Jjmj

32

2 Comments

AI is reshaping how enterprise software delivers—and prices—value. Companies like Zendesk, Intercom, and Forethought are shifting to outcome-based pricing, charging only when AI autonomously resolves issues. I've been investing in AI for over a decade. The shift to outcome-based pricing in enterprise software is a seismic change. As I told Jon Victor at The Information, software companies now face a critical decision: Does a company decide to change business models at the expense of losing revenue? It's a necessary risk. With cheap access to LLMs, businesses can build their own enterprise apps. Adapt or die. Why this pricing shift matters: → Vendor success ties directly to customer outcomes → Short-term revenue risks enable long-term market expansion → Industry faces widespread pressure to demonstrate measurable value Pitching AI-powered automation software as a clear cost-reduction tool could ultimately lead to greater revenue for vendors. This is because customers may view AI solutions as fundamentally different from traditional enterprise apps, whose business value is often harder to quantify directly. It's a huge market expansion opportunity. Vendors that nail outcome-based pricing could see massive growth as AI transforms enterprise operations. Those that don't? 😬🫣 How do you see this impacting sectors beyond customer support? Leave a comment. And if you're building in this space, send me a message.

35

3 Comments

In a recent interview, I was asked about the declining interest in strategy games over the last decade. Quantic Foundry adds data to the conversation but can't answer the fundamental question. They pose these, instead: 💡 Why do you think gamers have become less interested in strategic thinking and planning? Are there other potential causes that come to mind? 💡 Do you feel your attention span and ability to think deeply has changed since the emergence of social media and/or smartphones? 💡 Are there other documented changes in our media-consumption habits or cognitive metrics that you think might be related to this? 💡 Have you seen similar or related findings specific to gamers and games research?

11

15 Comments

Was chatting with another investor today on a call, and mentioned Tim Urban's Wait But Why Blog post, "The AI Revolution" and how prescient it was. Given that it was written in 2015, decided to flex AI's latest muscles to give it a new venue: a podcast. Podcast here: https://lnkd.in/gW8Dh89U Pretty easy to do: Step 1: Save the posts as PDFs. Step 2: Upload to NotebookLM, have it create a podcast. Step 3: Upload to Descript, give it captions. Step 4: Export, publish to Youtube. Total time to create the podcast: ~25 minutes. To generate this podcast and publish it even a few years ago would have taken >8 hours. Love the future (Original posts: https://lnkd.in/gW-PQXE7, https://lnkd.in/gg-nRvxz) #AIRevolution #GenerativeAI #WaitButWhy #TimUrban #NotebookLM #AIContentCreation #Podcasting #Descript #ArtificialIntelligence #ContentFlex #ProductivityHacks #FutureOfAI #AIinMinutes #TechEfficiency #AIandCreativity #YouTubePodcast #InnovationInContent #vibecap #vibecapital

5

Worth reading: Generative AI is not going to build your engineering team for you. So, so good - peppered with gem quotes and observations- chuckle. https://lnkd.in/gZBKtU4X Thanks to Eric Brooke for sharing.

16

Testing AI for ethical harms at scale for realistic failure modes is notoriously difficult and challenging. At vijil we did that, evaluating Llama 3 models vs their Llama 2 counterparts for harmful stereotypes. The outcomes were interesting. Check out the blog post to know how!

23

4 Comments

If you’re a founder building at the intersection of Games & AI, or just interested in what’s happening at the bleeding edge of the future in Silicon Valley, the new episode of /all chat is worth a listen.👇 Link in comments below. Hot takes from Patron portfolio founders Rick Ling of Rascal Games and Nico Christie of Altera with my partners Brian Cho and Jason Yeh! #games #ai #founders #seedfunding #startups

28

3 Comments