Raymond H.

"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

A lot of folks might be starting to think about curating festivals, exhibitions or other art things in 2025, so: if you’re interested in critical and nuanced thinking around AI, I have a new work for 2025 that might be of interest. “Human Movie: Six Meditations on a Compression Algorithm” is a 35-minute video essay contrasting computational processes of diffusion models and the human metaphors used to describe them — such as temperature, creativity, image recognition, memory, reason, and the unmodeled. Created from a blend of glitched AI-generated video, archival and found footage, the film is not about machines at all, but rather, seeks to assert a humanist counterfactual to comparisons between human thought and the limited capacities of generative AI. More at the link below, happy to offer previews to folks if you’re interested. #aiart #aicinema #aivideo #filmfestivals #criticalmedia

24

1 Comment

Ethical Considerations in AI Pricing: Balancing Profit and Accessibility by Edward Wong from Ibbaka What ethical issues come up when one uses AI to inform pricing? As adoption of generative AI based applications starts to take off it is time to put in place plans to address and mitigate ethical issues such as transparency, bias, availability and environmental impact. This is part of an AI pricing strategy. https://lnkd.in/g7ZaWJmm #ethics #pricing #AI

1

Exciting update: I will be reviewing AI products in detail, no BS, with real world scenarios, raw, for real. 🫡 🫶 Follow to stay informed. First: @Magnific_AI Full review: https://lnkd.in/gbh5h4Wg Exceptional Quality, Slow Processing, and High Costs for Creative Image Transformation Onboarding 🚀 Pros: - Fast Sign-Up: The sign-up process is quick and requires minimal steps. - Immediate Access: After sign-up, users are directly introduced to the pricing plans, allowing for quick initiation if they choose to subscribe. Cons: - Pay-to-Play: Users must subscribe and pay upfront to begin using the tool, which can be a deterrent. - No Free Sandbox: There's no free trial or sandbox mode to test the product before committing financially. Quality of Output 🎨 Evaluations: - Accuracy: The accuracy of upscaling images is mixed. For high-quality images, the output is impressive, but for lower-quality images, the results can be unpredictable. - Creativity: The AI shows a level of creativity but can go overboard, altering critical image elements. - Overall Quality: When the criteria are met, the upscaling and image transformation are highly detailed and sharp. Speed of Generation ⏱️ Time Measurements: - Initial Uptime: Starting the process takes between 1 to 2 minutes. - Completion: Total time for upscaling is approximately 5 minutes for each high-resolution request. - Efficiency: Compared to other AI tools, #Magnificai is relatively slower. Cost of Use 💰 Pricing Models: - Pro Plan: $39/month - Premium Plan: $90/month - Business Plan: $299/month - Free Trial: 2 months, with the option to cancel at any time. - Hidden Fees: Tokens/credits used for each operation can add up, increasing overall costs. Customization and Flexibility 🔧 Assessments: - Prompt Customization: Offers some capabilities to adjust settings, including HDR intensity and creativity calibration. - Model Fine-Tuning: Limited options available and every change costs tokens and consumes additional time. - Adaptability: Relatively low as exploring different configurations is resource-intensive.e Language and Multilingual Capabilities 🌍 Proficiency: - Language Support: Limited to English with no apparent support for other languages. - Non-Standard Dialects: No evidence of capabilities in handling regional dialects or idiomatic expressions. Collaboration Features 🤝 Features: - Multi-User Access: No features supporting multi-user access or team collaboration. - Real-time Collaboration: No real-time collaboration features are evident. - Shared Projects: Lacks functionality for sharing projects within teams. User Feedback Mechanisms 🗣️ Features: - Feedback Collection: No built-in mechanisms for collecting user feedback on the platform. - Satisfaction Adjustment: No features for rating or adjusting outputs based on user feedback. Report created by Pactto's AI 🔥 Have you played with it? Thoughts?

13

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

12

From where I sit in the SF Bay Area, cradle of innovation for at least the last 30 years, I'm a true believer in the promise of the #AI revolution. That said, I find it fascinating to consider views on the opposing side. I'm sharing below two opinion pieces from the last two weeks suggesting that the AI revolution is fundamentally overhyped and a bubble ripe for bursting, and that AI tools are actually not useful or transformative. Contrast this with venture capitalist Kai-Fu Lee's prognostication that 50% of white collar jobs will be displaced by #AI within three years, and that OpenAI and the other cutting edge movers in the space are headed to a trillion+ dollar valuation. Which side will win out? Where will we be by 2030 in the AI race? I'd love to hear your thoughts in the comments.

7

4 Comments

Meta's Llama 3 dropped just a few days ago, but already there are close to 1,000 variants publicly available on Hugging Face. Most people read base model evaluations and think they apply to all these deployments, but that ain't the case. Or at least the extent to which it's true is poorly understood. I want to see more eval work focus on the performance and risk shadow cast by base models on their downstream progeny. Basically, if base models aren't being used as-is, do their evals matter? Or are they poorly predictive of what will come of them post-fine-tuning? I don't have time/resources to work on this right now, but I'd love to see someone take this project on: across all variants posted on HF, how do evals on fine-tuned versions compare to evals done on the original base model? What kind of relationship is there between the two? Is the latter a constraint or directional anchor for the former? The best set-up here needs to find some way of ignoring some amount of noise (e.g., devs creating their own copy but not doing any fine tuning or doing some kind of terrible fine tuning) in order to focus on more serious attempts to fine tune for stronger performance on specific use cases, stronger safety and security, or something else. This question is increasingly important because fine-tuning is becoming easier and faster to do. That means it's more and more likely that developers will tune models to their own liking. Evals done on prior upstream versions are potentially meaningless... 'potentially' is the key word because we don't have a good understanding of the persistence of performance and risk, the elasticity of learning. By the way, if you're interested in working on hard questions around LLM evaluation, like these ones, we're hiring! Our Trust & Governance research team is looking for a Staff Applied Research Scientist. Come help us tackle these challenges! Link to the job posting in the comments. #artificialintelligence #trustworthyai #llm

51

9 Comments

Some context from the #Founder perspective. As liberal as Seattle is as a city with policy and politics, it's incredibly conservative from an investment standpoint. One of the reasons you may not hear big "home-run" type pitches or pitches with sizzle, is simply because investors in Seattle don't like those types of pitches and Founders have been ingrained to not present in that manner. They may say they do, but ask any Founder and they'll tell you how they have been told to size down their pitch or talk about MVPs, first customer, first revenue, etc. or how what matters more is what you can do first couple years rather than the home run. Those early metrics are important by the way, but different from the big home run. Do you know who is receptive to homerun pitches? Investors in the Bay, NYC, Miami, Hong Kong. Seattle is not great for early-stage Founders with many prioritizing raising outside the city. We will be raising capital in the fall and as of now, have no plans of reaching out to Seattle-based VCs. I should also add that our company has a blockchain component and sadly Seattle VCs have not embraced crypto or blockchain. Welcome to #Seattle, Kyle Lui! Hopefully, you can help change the narrative! GeekWire

22

7 Comments

Was a bit surprised to see Wired touting RAG as "taking hold in Silicon Valley" like it's the hot new thing - RAG has been around for ages (at least in LLM terms) and I actually think its application space is currently shrinking as context windows grow. Their "reduce AI hallucinations" thing is not really the whole story - I mean, yes, RAG can definitely help in the sense that asking a broad question to the base model (with no input other than the prompt) is likely to result in hallucinations, whereas feeding in concrete data (with RAG or otherwise) will ground the response in that data and therefore produce fewer hallucinations. But RAG is really just a tool for feeding that data - and, yes, for massive search tasks (like broad legal research requests across entire databases), RAG may be the best (if not only) way to do it. But for smaller tasks (like analyzing a single document) it's often better to just feed the input in directly, which the larger context windows of the newest models increasingly allow. I feel like a few months ago (around the time Opus was released) people were asking whether RAG would be obsolete - that was an overreaction, but still it doesn't really fit this narrative of RAG as some hot new thing in the LLM world. Maybe I missed the pendulum swinging the other way again? Not sure. https://lnkd.in/eUEwBmFf

3

Today marks an occasion, Dan Luo wrote her first blog post for Ava Tech! Head on over and have a read, hopefully you will learn some cool stuff about expressions in Make! And remember if you are curious about what Make is or how you might be able to use it in your everyday life or business, flick us a message! #automation #integration #WorkflowOptimization #AdvancedExpressions #MakeAutomation https://lnkd.in/gerQxfSH

7

The predictable enshitification of maps with ads: https://lnkd.in/g8ptc4_s But hey, what if it was all more contextually relevant? Like what if you really needed gas and it showed you gas stations nearby? Isn't that useful? That's still a very likely "no" -- esp. if the ad-tech company is using your personal information against your personal best interests. E.g., even if you do need gas, the system can hurt you if the ad auction winner is more expensive, got worse reviews, and was farther out of your way than the other options you weren't told about. FWIW, at Keyhole (which became Google Earth after acquisition) we also played with how to monetize maps via ads. Serving earth data is very expensive, esp. for a small company without cloud resources. The best idea we had was to promote map labels to show up earlier or more prominently when zooming or panning in, based on a prior auction. Turns out, you never want to show all possible map labels at once -- TMI. Highlighting a key city name often leverages negative space (no labels) around it. So the ad auction would be one way to pick and prioritize which comes first to make everything more readable. Contextual picking is even better. We also played with less conspicuous visual ads. But none of these were very satisfying. All felt kind of icky, so we never did it. Google won our attention in the 2000s not simply because search worked better than the other options, but because the text ads were small and easy to ignore. Remember that? Ad-tech business models always drive to the worst outcome over time.

8

I just gave Google’s NotebookLM my entire book (co-authored with Clayton Pummill), and what came back was a surprisingly good way to analyze a document. If you're unaware, NotebookLM turns books and documents into an AI-powered NPR-style podcast plus a study guide, FAQ, timeline, and a surprisingly accurate chat experience—all in one. The podcast version is fun and provides a compact dive into the book’s themes. That said, it skims the surface and doesn't completely capture the book's nuance. Still, neither do most podcasts due to the temporal constraints and the limitations of an author's ability to cover the same information verbally as in writing. It’s tough to match the depth and precision of writing reflected in painstaking rewrites—after all, that’s what writing’s for, right? What also stood out was how the “hosts” used clever tricks to keep the conversation going. It all feels very organic. I thought it would feel contrived and give me flashbacks to the Rogerian tricks ELIZA used. Perhaps this format will become grating after more iterations, but my first impression is that NotebookLM entertainingly captures the core ideas and does so in an engaging manner. I will certainly be trying it on some ARVIX papers that I have not had the time to read. You can check it out here:

30

10 Comments

SB 1047 is a pending California law that is essentially outlawing large open source models. If you like open source (and I do!) this would be very bad. And now it is reported that one of the advocates of the bill is trying to cash in on it 🤦‍♂️ The bill technically doesn't outlaw large open source models, but it requires "PWC type audits" costing "on the order of hundreds of thousands of dollars to single digit millions." As most open source authors or university research labs don't have that much money, it achieves the same outcome. It would also make AI start-ups vastly more difficult and expensive. The AI community (or at least the part of the AI community that likes open source) unsurprisingly are trying to stop it. Today, the article below reports that "Dan Hendrycks, an executive at Center for AI Safety — the firm whose lobbying arm helped draft the bill — co-founded a company called Gray Swan that offers AI safety compliance tools that seem positioned to provide the type of auditing data the bill would require". https://lnkd.in/gdutbeKq

64

14 Comments

I'm thrilled to see Anthropic's latest announcement of the upgraded Claude 3.5 Sonnet and new Claude 3.5 Haiku models! 🚀 As someone who strategically leverages various AI platforms, including Claude, for specific use cases, these advancements are particularly exciting. The enhanced coding capabilities in 3.5 Sonnet are especially promising 💻✨, as they build on Claude's already impressive performance in this area. What really stands out to me is the groundbreaking computer use capability now in beta. 🤖🖱️ While I currently use AI selectively and thoughtfully for specific tasks, features like this - where AI can interact with computers like humans do - hint at the expanding possibilities ahead. I appreciate Anthropic's transparent approach in acknowledging that this feature is still experimental and may have limitations. This aligns with my philosophy of carefully choosing when and how to implement AI tools as they mature. 🎯 The fact that Claude 3.5 Haiku matches Claude 3 Opus's performance while maintaining cost efficiency is also noteworthy. 📊💪 This kind of advancement helps organizations like ours make informed decisions about scaling our AI usage as the technology evolves. Looking forward to exploring these new capabilities! 🌈 #AnthropicAI #Claude35 #AIInnovation #ComputerUse #FutureOfTech

2

I've recently been learning about using distance metrics (specifically cosine and Euclidean distance) to determine the similarity between multidimensional arrays of vectors. "Very nice", I hear you say, "but why would you want to do that?" If you pass two images to a convolutional neural network that has been designed and trained to extract feature representations from images, those feature representations will be encoded as multidimensional arrays of vectors. Once you have a pair of feature representations from two different images, you can use a distance metric to compute the distance between those feature representations. You will need to decide on a distance threshold. If the distance between the two feature representations is less than the threshold, the images are classed as being the same and if the distance is greater than the threshold the images are classed as being different.

7

Black Forest Labs launches FLUX.1: a new image generation model for generative artificial intelligence - FLUX.1 suite of generative models launched - $31 million raised in Series Seed funding round - Models accessible for diverse applications: commercial, non-commercial and personal - Advisory board with industry experts announced ps. the enclosed image has been created with the new model. Please upvote and share. Read more on our magazine by clicking below 👇 #flux #blackforest #genAI #imagegeneration #aiimages #generativeAI #comfyUI https://lnkd.in/dTHMb27u

11