Dali Kilani

➰ APIContext's latest insights on cloud-based API performance have been featured on Nordic APIs! In the 2024 report, we delve into crucial aspects affecting API efficiency: 🔐 Authentication is key. Ensuring robust authentication mechanisms in APIs is crucial, yet offloading this function can introduce complex challenges. 📖 DNS configurations matter. How you handle DNS settings can significantly impact API performance, particularly when tied to authentication processes. 🌐 The role of third parties. Critical third-party services play a significant role in 🔗 API transactions, underscoring the need for careful integration and monitoring. ⛅ Not all clouds are created equal. Differences in cloud services can drastically affect API performance, an essential consideration as APIs account for over 80% of all internet traffic. #APIs #CloudComputing #TechInsights #APIContext

15

1 Comment

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

I had the luxury to wake up in the sunny alps this morning! Yesterday, I enjoyed some great #AI in the #Alps real talk and discussions: 1️⃣ On AI certifications: Why sampling the application domain (a.k.a. constructing the test set) is the crux of the matter and which methods can be used to approach that systematically on a meta level. 2️⃣ On AI-based simulation: That the next generation surrogate models are LLM-based, much more powerful and probably won't use any a-priori physics based constraints. 3️⃣ On leadership and change management: Why German organisations tend to make slow ('ich muss da erst nochmal mitnehmen') and timid ('ich muss da erst noch jemanden abholen') decisions. + some great life advice in the evening 😉 Thanks for organizing this event and making it memorable! Robert Weber Peter Seeberg Barbara März Johannes Brandstetter Bernhard Nessler Klaas Nebuhr Dr. Andreas Nauerz Michael Fischer Kirsten Prill Bastiaan O. Werner Faulhaber Fabian Kreutmayr + many more... #ai #alps #collaboration #realtalk

80

7 Comments

Bowen Yang from Cohere will dive into the world of long-context models and how we can push the boundaries of transformer architectures. At TMLS2024 we'll explore: ✅ Scaling Transformers: Techniques for handling longer inputs while maintaining efficiency. ✅ Challenges & Solutions: Examining the modelling and framework hurdles we face, along with potential solutions. ✅ A Range of Approaches: We'll delve into methods across data, modelling, and framework levels. ✅ Training, Evaluation, & Inference: Covering the entire NLP pipeline for long context tasks. This panel promises to be a valuable resource for anyone interested in pushing the limits of language understanding. See the full abstract here: https://lnkd.in/gwSRjPjz #NLP #Transformers #LongContext #MachineLearning Toronto Machine Learning Society (TMLS)

30

💡 Why do we need so many evaluation tools for LLMs? As engineers, we build 'production-ready' LLM products using these metrics. But what happens next? How do we maintain control and ensure reliability? At Feedback Intelligence, we’ve crafted a cookbook to keep your LLMs reliable and aligned with user expectations. 🍲 📖 Give it a read and let’s chat!

16

🌟 Introducing Nexizo - https://Nexizo.AI 🌟     When we launched BidAssist our vision was to empower businesses to query public procurement data, aka government tenders, with the ultimate goal of helping them grow (a given, being an OfBusiness product) Over the past few years, our team has organized, normalized, and categorized over 50 million tenders from 80,000+ sources into a single schema using AI/ML to be able to provide a queryable tender data feed to few very select beta customers👌. This journey has helped us gain deep insight into how customers leverage this data for market intelligence, and to identify prospective customers. Most customers have seen significant growth upside (2X-4X) 📈. While doing this, we also stumbled on another learning ✅ Most customers find it challenging to query or utilise government data to its full potential when left to themselves. They tend to get lost in the millions of potential data points that can be used for decision making. However when we worked very closely with the customers, it helped them unlock exponential value from our #GovTech #AI stack. Kudos to our solutions team 🫡 The approach not only helped settle the apprehensions of yet another over promising #AI / #SaaS product but also helped find more & more use cases to grow the ecosystem. With these learnings, we launched Nexizo.AI 🚀    Simply put, it’s designed to provide solutions that extract actionable insights from public procurement data, tailored to your needs. Our platform stands out by helping businesses uncover relevant B2B opportunities enriched with critical contact and financial information. For the curious ones 💡  >> BidAssist remains the go-to search engine for tenders; while Nexizo, inspired by Oxyzo Financial Services promises to scale your business with next-generation technologies On a daily basis, we process 50K tenders/day from 50+ languages.  More on Nexizo 👇 1. **Unmatched Data Insights**: Our AI algorithms efficiently analyze vast amounts of data, identifying the most promising prospects  2. **Enhanced Decision-Making**: Access actionable insights to make strategic, informed decisions with confidence.   3. **Streamlined Business Expansion**: Simplify the process of finding and connecting with relevant B2B companies to drive With Bhuvan Satya Nikhil Maheshwari Nikunj Parikh Nishant Jadon Atul Sharma Avichal Sahai Sameer Ahmed Shivam Gupta Rahul Bansal Chandna Gupta Follow us for updates and success stories. OfBusiness Nexizo Oxyzo Financial Services #B2B #AI #BusinessGrowth #NexizoAI #OfBusiness #Innovation #GovTech #DataDriven #NextBigThing

126

3 Comments

First morning of GenAI Summit highlights: From Lukas Biewald talk on Evaluation - while there is not a common quality benchmark to AI quality, it is critical to define your own and build it right after your PoC and way before moving to production - when working with AI, keeping records of your experiments, both failed and successful, is probably the most valuable part of your IP. That is counterintuitive to most SW development teams, used to only "commit to their repo" the "working code". From Tengyu Ma talk state of RAG - the 2023 discussion between long-context vs fine-tuning vs RAG is finally over, and RAG is the clear way to go - in the near future Embedding models and LLMs will handle most of the work (like multimodal embeddings generation), and we'll be able to remove many of the manual tricks that we need today to make RAG accurate - privacy = on-premise 👀

15

3 Comments

🚨I'm honored to be part of this year's FF PMC. Flink Forward is the place, where the #apacheflink community gets together every year, organized by Ververica | Original creators of Apache Flink®. Why Flink Forward? 👉 #apacheflink although around for years, has exploded in demand over the last ~2-3 years. And there is a good reason for it 😉 👉 Due to its nature and use from many early adopters and hyperscalers, you get to see the "what's and how's" of (but not limited to) #streaming problems; the biggest names in the market aim to solve. 👉 What's the next big thing for #streaming and #streamprocessing systems. ... and way way more. If you thought last year was a big show (even though we got our drones 🛫 jammed 😀 🙃 ), then you are up for a surprise as this year will be HUGE. It's Flink's 10th anniversary 🎉 , so it's only natural cool stuff is in place 😎 🌊 If you have a story to share and haven't done already make sure to submit your proposal: https://lnkd.in/dQjApM_d 🌊 If you feel reluctant or want more feedback, you can always reach out to any PMC; we are here to help 😉 🌊 Let's rock these streams 🤘 and look forward to seeing you all in Berlin. #dataengineering

52

So - are LLMs Universal? This guy looked at a lot of proofs that LLMs are universal and concluded that they are all wrong or misleading. I didn't check all the proofs and I didn't check all the errors or misunderstandings that he mentions. But I would say that it is not confirmed that any realistic model of LLMs is universal. Specifically about https://lnkd.in/da4qHFgh I did not make up my mind if the model is realistic or not, but I think it makes improvement by 1) Getting the outer while loop into the LLM by emphasizing auto-regression 2) Getting the ability to skip parts of the input with the bounded context window, whether a realistic aspect of a theoretical model or not.

4

1 Comment

Looking at this on another axis: with any new technology, we begin by trying to make it fit the problems we already have, while the incumbents try to make it a feature (hence Google and Microsoft spraying LLMs all over their products in the last year). Then startups use it to unbundle the incumbents (to unbundle search, Oracle or Email), but meanwhile, other startups try to work out what we could build that would be truly native to the new technology. That comes in stages. First, Flickr had an iPhone app, but then Instagram used the smartphone camera, and used local computing to add filters, and further on again, Snap and TikTok used the touch screen, video and location to make something truly native to the platform. So, what native experiences do we build with this, that aren’t the chatbot itself, or where the ‘error rate’ doesn’t matter, but abstracts this new capability in some way? // https://bit.ly/3yZVU5m

6

Sam Altman has declared that we are only a few thousand days away from 𝘀𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲. (blog link in comments) According to ChatGPT (why not ask the source!?!), 𝘀𝘂𝗽𝗲𝗿𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 is "𝘢 𝘧𝘰𝘳𝘮 𝘰𝘧 𝘪𝘯𝘵𝘦𝘭𝘭𝘪𝘨𝘦𝘯𝘤𝘦 𝘵𝘩𝘢𝘵 𝘴𝘶𝘳𝘱𝘢𝘴𝘴𝘦𝘴 𝘵𝘩𝘦 𝘤𝘢𝘱𝘢𝘣𝘪𝘭𝘪𝘵𝘪𝘦𝘴 𝘰𝘧 𝘵𝘩𝘦 𝘴𝘮𝘢𝘳𝘵𝘦𝘴𝘵 𝘩𝘶𝘮𝘢𝘯𝘴 𝘪𝘯 𝘷𝘪𝘳𝘵𝘶𝘢𝘭𝘭𝘺 𝘦𝘷𝘦𝘳𝘺 𝘥𝘰𝘮𝘢𝘪𝘯." I work in the 𝗯𝗮𝗻𝗸𝗶𝗻𝗴 domain, and a year ago, I'd say that ChatGPT - and its underlying models - were the equivalent of recent college grads, trying to pass the bar exam. Newbies, to be frank...and if I'm really honest, they were surprising bad at simple math. Nevertheless, what they were good at -- general knowledge -- was impressive and had real utility! Fast forward to now and the foundational models are well beyond basic utilty; they are now 𝘷𝘦𝘵𝘦𝘳𝘢𝘯𝘴 of industry! Not only are they great at math now (yes, they show their work!), they are even great at teaching math! Furthermore, the latest FMs demonstrate a remarkable mastery in the dark recesses of loan underwriting best practices of niche financial products! 𝗗𝗼𝗻'𝘁 𝗯𝗲𝗹𝗶𝗲𝘃𝗲 𝗺𝗲? 𝗧𝗿𝘆 𝘁𝗵𝗶𝘀... Ask ChatGPT about the eligibility requirements for a USDA B&I loan. Don't spell out what "USDA" nor "B&I" stand for...just ask it like you would ask a veteran commercial banker. Let me know what you think about the results! P.S. If you're interested in exploring the limits of how AI can aid humans in commercial lending, give my friend Michael Marcucci, CRC, CIA a follow. He's been on the bleeding edge of exploring this for a few months, and has been documenting his findings on LinkedIn! P.P.S. If you want to 10X-100X your lending efficiency by putting these types of capabilities to work for your business, head over to Cascading AI (YC S23) and book a demo! #banking #financialservices #artificialintelligence #ai #genai #superintelligence #openai #samaltman

11

3 Comments

🤖 Creating a fully autonomous AI developer to write and maintain code is still the holy grail of software development. There are many brilliant teams currently working on this, and each of them have their own unique approach. The team at Cosine have made a massive step towards making this a reality. “Cosine achieves state-of-the-art results on the SWE-bench benchmark” They have achieved a score of 43.8% on SWE Bench, which was verified by OpenAI. What's even more incredible is they have managed to do this with a lean team and a relatively limited budget. Though it’s important to note Cosine isn’t looking to replace human developers entirely, but rather augmented them with human-like assistants that they can collaborate with on any kind of coding task. Yang Li + Sam Stenner + Alistair Pullen Read more about their collaboration with OpenAI 👇 https://lnkd.in/gu7xXVPn

21

🚀 Want to run your very own trained LLM, privately and securely? 🤖🔒 Imagine harnessing the power of a Language Model tailored just for you without worrying about third-party servers! 💼✨ Try, LM Studio. Jest for few click . LLM running in your PC. With LM Studio, you can ... 🤖 • Run LLMs on your laptop, entirely offline 📚 • Chat with your local documents (new in 0.3) 👾 • Use models through the in-app Chat UI or an OpenAI compatible local server 📂 • Download any compatible model files from Hugging Face 🤗 repositories 🔭 • Discover new & noteworthy LLMs right inside the app's Discover page #LLM #private #chatgpt #AI

9

If you’re developing with AI/ML follow Julian Norton and Plum They’ve got a top-tier grip on the market and know how to increase model efficiency significantly. Check them out

15

1 Comment

Great to see not only one but two of our portfolio companies in a podcast! If you want to learn how you can squeeze double (or more) the performance out of your database, have a look.

25

1 Comment

"All AI can do is semi-automatic screensavers", said legendary director Benicio Del Toro. And indeed, when you watch this video I just generated with Kling, it seems clunky, dumb, superficial, almost hopeless. But Benicio is wrong. The progress in AI video is mind boggling. In a few years, these kind of early AI AI-generated videos will feel like the mute, black and white videos from the 19th century - they were clunky, but they were the first brick of what would become a foundational industry. It has been said time and again: when photography became popular, painters refused to called it art because it was "low-effort". Now, directors that have been through the nightmares of production (and I know what I'm talking about) consider AI "low-effort", and as such, not worthy of their attention. Ironic, given how many of them use CGI instead of practical effects. Months ago, we didn't believe AI could comprehend any physics. Now, it's getting pretty good - you can even help it understand what should move. We can foresee the next models getting even better at that, with consistent characters, places, vehicles, etc. Already, millions of new creators are painstakingly creating first video concepts with the very clunky tools at their disposal. And it's only going to get better. I'm extremely excited at the idea of being able to one day create a movie out of my science-fiction novel, without having to spend millions of dollars. I will do it with the passion of artistic craft that I used to create my movie - the medium will just be different. AI will create millions, if not billions of new artists and creators. Let's embrace the technology.

33

2 Comments

For anyone interested in the question "How much better than GPT-4 will AI become?" — I strongly recommend reading fascinating essay (link below) on the future of AI by Leopold Aschenbrenner, who recently left OpenAI's superalignment team. --- You can see the future first in San Francisco. Over the past year, the talk of the town has shifted from $10 billion compute clusters to $100 billion clusters to trillion-dollar clusters. Every six months another zero is added to the boardroom plans. Behind the scenes, there’s a fierce scramble to secure every power contract still available for the rest of the decade, every voltage transformer that can possibly be procured. American big business is gearing up to pour trillions of dollars into a long-unseen mobilization of American industrial might. By the end of the decade, American electricity production will have grown tens of percent; from the shale fields of Pennsylvania to the solar farms of Nevada, hundreds of millions of GPUs will hum. The AGI race has begun. We are building machines that can think and reason. By 2025/26, these machines will outpace college graduates. By the end of the decade, they will be smarter than you or I; we will have superintelligence, in the true sense of the word. Along the way, national security forces not seen in half a century will be unleashed, and before long, The Project will be on. If we’re lucky, we’ll be in an all-out race with the CCP; if we’re unlucky, an all-out war. Everyone is now talking about AI, but few have the faintest glimmer of what is about to hit them. Nvidia analysts still think 2024 might be close to the peak. Mainstream pundits are stuck on the willful blindness of “it’s just predicting the next word”. They see only hype and business-as-usual; at most they entertain another internet-scale technological change. Before long, the world will wake up. But right now, there are perhaps a few hundred people, most of them in San Francisco and the AI labs, that have situational awareness. Through whatever peculiar forces of fate, I have found myself amongst them. A few years ago, these people were derided as crazy—but they trusted the trendlines, which allowed them to correctly predict the AI advances of the past few years. Whether these people are also right about the next few years remains to be seen. But these are very smart people—the smartest people I have ever met—and they are the ones building this technology. Perhaps they will be an odd footnote in history, or perhaps they will go down in history like Szilard and Oppenheimer and Teller. If they are seeing the future even close to correctly, we are in for a wild ride. Let me tell you what we see. --- What are your thoughts on this? ♻️ Found this useful? Repost for your network.  🏡 Visit my profile to fully automate admin work in HubSpot

16

10 Comments

Every time we sit down with enterprise CIOs and AI teams, the top of mind concern we hear is "Reliability" - how do we go from our use case to reliable AI as fast as we can? In our experience, reliable AI is extremely hard to engineer. There is no silver bullet - it is about the many long tail problems that you need to grind at for weeks and months. But, this is what makes enterprise AI extremely resource intensive today. We are building Awen to change this. We are building the core agentic infrastructure that helps AI teams ship reliable AI 100x faster - from data curation to running evals at scale. Experience more in our playground: https://lnkd.in/gFGnpcdR Abhishek Borar Himanshu Muchhal

9

Building in Public | #DigitalNomads Silly thought experiment. workforce automation becomes a thing. workforce automation tools are built on LLMs. cost of compute trends towards the cost of energy companies will seek lower cost "digital employees" and will "outsource" to low energy cost countries.

10

9 Comments