Tessa Lau

You don’t have to live in California to understand the implications of proposed California regulation SB 1407. Take a moment and read Andrew Ng’s latest post highlighting why this regulation is the wrong way to pursue AI safety.

10

💥 $1.6 𝑩𝑰𝑳𝑳𝑰𝑶𝑵 𝑫𝑶𝑳𝑳𝑨𝑹 𝑷𝑫𝑭?!! 💥📚 𝘛𝘩𝘦 𝘊𝘢𝘭𝘪𝘧𝘰𝘳𝘯𝘪𝘢 𝘏𝘪𝘨𝘩-𝘚𝘱𝘦𝘦𝘥 𝘙𝘢𝘪𝘭 𝘌𝘯𝘷𝘪𝘳𝘰𝘯𝘮𝘦𝘯𝘵𝘢𝘭 𝘐𝘮𝘱𝘢𝘤𝘵 𝘚𝘵𝘢𝘵𝘦𝘮𝘦𝘯𝘵 This report is so massive (and expensive) that the San Francisco Public Library had to wheel it out on yellow trolley for me… and this was only Chapter One!! 📚🛒 When we talk infrastructure and buildings, most people don’t realize the vast quantity of regulatory paperwork and consultants required to get construction projects started. The environmental reviews, approvals, and impact assessments are not just formalities—they’re a mountain of regulatory work that can take decades to complete and cost millions… or in this case, 𝐛𝐢𝐥𝐥𝐢𝐨𝐧𝐬 𝐨𝐟 𝐭𝐚𝐱𝐩𝐚𝐲𝐞𝐫 𝐝𝐨𝐥𝐥𝐚𝐫𝐬. Is this the most expensive document ever written? 🤔 #infrastructure #regulatorycompliance #environmentalreview #californiahighspeedrail #buildingsustainability #transportation

30

4 Comments

There you have it! BREAKING NEWS! New model comparison feature up on all Ask Sage, Inc. tenants! This feature enables you to compare prompt results across LLMs on a single screen with RAG, live, file attachment options. Did you know we can support any commercial and OSS models and deploy on any environment (air gapped, cloud, on premise... ?) Great convenient feature!

111

7 Comments

The best AIs will be constructed not emergent.  https://lnkd.in/gMTah5ns Some predictions on the next era of AIs with Eric Schmidt. We argue that the best AIs will be constructed (purpose built) rather than purely emergent. General purpose foundation models will serve as important infrastructure (not applications) with close tie-ins to cloud and consumer applications. AI training will be a key differentiator between the best AIs and the rest, with prompting being  “flashlight” (versatile but less powerful + precise) and training a “laser beam” (powerful, precisely controllable). This is especially important as we transition to AI agents (still in the future!), which can tackle entire workflows. But in doing so we face challenges with compounding errors, which demand the most performant AI capabilities. The real world has a combinatorial explosion of workflows, and we think it unlikely that they consolidate around a single future foundation model. We outline a “brick by brick” AI construction that has complex capabilities calling simpler capabilities like subroutines. We see this structure in many of the powerful AIs today, with more such to be built in the future.

128

5 Comments

The harsh reality of CA politics -- the state government works for Big Tech, not the citizens. Some good quotes here. Newsom's stated goal about workable "guardrails" is nonsensical as it's currently physically impossible. That's a bit like saying they are pursuing a way to apply DDT safely to agriculture.... no such method existed. Newsom and CA have zero cred on anything relating to responsible or safe AI. The question is whether that extends to Harris. I honestly don't know, but her history in defending SV and conflicting relationships that include her brother-in-law suggests that it would be difficult for her to regulate Big Tech objectively. Trump is also a wild card. He's so unpredictable that one day he might attempt to order Zuch arrested and the next brag about the power of American Big Tech. The next four years will be interesting. I have much more confidence in SCOTUS.

6

5 Comments

Notion vs Glean reminds me a lot of Tesla vs Waymo... Just like I think Tesla is going to solve the 'software' problem of autonomous vehicles before Waywmo solves the 'hardware' problem... I think that Notion is going to solve 'enterprise knowledge search' before Glean solves the problem of document and knowledge creation... Why? A couple reasons: 1) Just like Tesla, Notion has a huge advantage when it comes to the amount of underlying data and documents to train on... They already have 100M+ users, and likely 10x that in terms of the number of documents, files, etc. in team's workspaces. 2) It's very hard to change people's behaviour re: where/how they create 'work'. It's why a lot of companies have come and gone in that category... As exciting as Glean's AI writing assistants might seem, most people are just fine (or even prefer) a good old fashioned editor/writer. I could be way off here, but this will be an intersting one to follow. Thoughts? Where am I getting it wrong? #ai #enterprise #knowledge #management #notion #glean

3

1 Comment

We need to talk about Agents! Inspired by LlamaIndex's recent tutorials on Agents (Ravi Theja Desetty) and Jerry Liu's talk on Advanced Context-Augmented LLM Applications, this post looks at the Planning component of an agentic workflow. We go through Parallelizable sub-queries, Task Decomposition and Planning with Feedback in detail to help LLMs "reason" vs "generate". Happy Reading! #ai #planning #agents #workflowautomation #tech

44

LLama3 passed the super-proprietary, advanced test that we designed for modern LLMs! almost... You can read about that test here: https://lnkd.in/gPG8WndR When I first saw the response generating, it looked like I was going to have to come back here and apologize for slandering these machines when they were just one iteration away from solving very basic prompt injection issues. But then I looked a bit closer... can you see the mistake? #promptinjection #countingishard #LLMs #LLama3

13

4 Comments

So many of you have reached out to me to get my thoughts on this. With the integration of powerful on-device AI, these new PCs promise to revolutionize how we interact with our machines. But as we peel back the layers, we find a more nuanced story – one of incremental advances, potential paradigm shifts, and significant implications for privacy and security. Read my full analysis here…

8

1 Comment

This is a good article on using LLMs in the real world. It covers many aspects that we learned at Nurish over the last few months and more. Using LLMs is an inexact science, but has its own needs in processes and tools. While most of the attention has been on the LLMS, their capabilities, and RAG, there are mundane issues that need to be taken care of, issues where the solutions for pre-LLM development are not that well-suited. These include version management, integration, testing, data management, quality assessment, and more. This article covers many of those practical aspects. Eugene Yan, Bryan Bischof, Charles Frye & more https://applied-llms.org/

9

Another great example of Seattle's tech ecosystem attracting investment, promoting growth, and potentially freeing up talent to either start their own new firms or attract new employers to enter the market to hire them. Transactions like this are part of the reason why Seattle remains one of the fastest growing regions in the country. #SeattleTech #TechInvestment #GrowthOpportunities #RealEstate

10

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

What if we could bridge the gap between what we see and what we say? This fascinating research paper delves into the exciting world of Vision-Language Models (VLMs)! Imagine a future where your visual assistant seamlessly guides you through a new city, or where AI conjures up images based on your wildest descriptions. VLMs hold immense potential to revolutionize our interaction with technology. But there's more to the story than meets the eye (pun intended). 👉 The paper dives into the key challenges: how can we translate the vast, continuous nature of visual data into the discrete world of language? 👉This comprehensive paper offers a roadmap for anyone curious about VLMs. 👉It unpacks the core concepts - what they are, how they work, and how we train them. 👉The researchers even explore methods to evaluate their accuracy, ensuring these models are reliable partners in our future. 👉And it doesn't stop at images! The paper also explores the exciting possibilities of extending VLMs to understand and generate video content. This research is a stepping stone towards a future where machines can truly see and understand our world. Read the paper here - https://lnkd.in/gHqNN7bJ

10

4 Comments

Just completed the entire Claude tutorial series from Anthropic! 🎓 https://lnkd.in/gq-3e93w. Very helpful to introduce in practical terms the use of artefacts and projects within Claude's front end. A huge shoutout to Elie Schoppik for the outstanding presentation and engaging teaching style. His clear explanations and practical examples made complex concepts around AI assistance truly accessible. The tutorials gave me a solid foundation for leveraging Claude's capabilities, from basic interactions to advanced use cases. Really excited to put these learnings into practice! 💡 Building competency with AI tools feels more crucial than ever, and resources like these make the journey both manageable and enjoyable. #AI #ContinuousLearning #Claude #Anthropic #TechSkills

2

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

This will unlock a lot of possibilities, and comes with some really nice advantages like: - More efficient chatbots - Easier builiding proces - Better customer experience - Cheaper prices for chatbots Now's the best time to start integrating customer support chatbots in your business to deal with customer while you focus on selling 💯

2

AI that can use computers like humans? Anthropic just made it a reality. Anthropic's latest announcement is reshaping the AI landscape: 1. Upgraded Claude 3.5 Sonnet with industry-leading software engineering skills 2. New Claude 3.5 Haiku model matching top-tier performance at lower cost and speed 3. Groundbreaking "computer use" capability allowing AI to navigate interfaces like humans 4. Significant improvements in coding, with Claude 3.5 Sonnet outperforming all publicly available models on SWE-bench Verified 5. Early adoption by tech giants for automating complex, multi-step processes This leap forward in AI capabilities opens up exciting possibilities for automation, software development, and open-ended research tasks. However, it also raises important questions about responsible deployment and potential risks. What are your thoughts on the implications of AI systems that can use computers like humans? How might this change the future of work and technology? #ArtificialIntelligence #TechInnovation #FutureOfWork #AIEthics #SoftwareEngineering Share your insights in the comments!

5

2 Comments