A. Kirimgeray Kirimli

New source alert! OmniAI (YC W24) now supports files as a source. Easily add any number of files and use LLM transforms to categorize, extract, summarize, etc. them. 🎉 Reach out if you're working with any unstructured data, including files!

68

Vizzly's changelog for May: 🔌 Plugin API / Custom Views • You can now inject custom views from third-party libraries into the Vizzly dashboard. 🇺🇸 New US region • We launched a new server in the US (us-east-2). Performance will always be a priority at Vizzly; we know there's nothing worse than waiting 419 years for a dashboard to load. ➗ New Custom Metrics available • Users can now calculate the difference between two date values and create new metrics based on conditions set before aggregation. 📆 Relative Filtering • When applying dashboard-level date filters, you can now set relative dates. For example, you can filter data for 4 months in the future or past. ✨ Other developments • ID Aliases • Value Aliasing • Redshift 'super' data type support • Snowflake array support ✌️

38

1 Comment

In November, we released a new product for SaaS companies called Census Embedded – it lets you embed our 200+ integrations directly into your application with a world-class API. We made our battle-tested connectors accessible to any developer. Since then we've helped companies like S&P Global, Metronome & Labelbox add these integrations to ingest/export billions of records. Today, we're making it easier than ever to implement with Embeddable Components. If the v1 was like Stripe's API, this launch is Stripe Checkout/Elements. Developing integrations is challenging because each API has varying requirements – different required fields, nuanced sequences of API calls, weird combos of CRUD operations, and more. We've spent years perfecting the transfer of data between 30+ data sources and 200+ destinations. We’ve solved scalability, reliability, observability, and UI quirks too! We’ve developed a sync and data mapping UI that accounts for the quirks in all 6,000+ combinations of those sources and destinations. Embeddable Components save you the hassle of reinventing the wheel in your own app. Offer your end users full customization of their data synchronization into and out of your app (or thoughtfully constrain those options).

97

6 Comments

#Fintech organizations have 3 main NEVERS: NEVER want to see downtime NEVER default on regulations NEVER miss a chance to know their customers But achieving this at scale = Ingesting TBs of data = spending millions of $$$ on SaaS observability tools = sharing financial/PII data with vendors = compliance issues This simple yet straightforward (+ extremely costly) gap This is where Incerto’s Observability solution comes in = Single unified platform, custom-built to your infrastructure = 85+% guaranteed cost savings + 100% control = #OwnYourObservability Still on the fence on making the switch from SaaS observability? You’d surely want to hear from our clients from FYERS, Dhan, and other leading orgs. #observability #financialservices #monitoring

22

If you're connected with me, you know I scream about how necessary integrations are. Check out this piece where I explain more for insideBIGDATA!

21

If you know what is stream processing, you probably should follow Michael Drogalis. If you are bootstrapping your business or in early stage startups, you should follow him on Linkedin. Michael, founder at ShadowTraffic, was the PM for Kafka Stream and ksqlDB, after his company was acquired by Confluent 2018. So glad to meet Michael Drogalis finally in person last week in sunny Seattle. In this episode of "Streaming Caffeine", Michael shared how to make live demo better with live data, how to win customers' trust, challenges and opportunity for stream processing, and AI of coz. Check the 20-min full version at YouTube 📺 https://lnkd.in/dgzcGx3Y

25

GM! Excellent post from my colleague at MIN3 - Katie H.. It is the latest in her series of video explainers covering some of the demo AI and crypto products that MIN3 has created to deliver real business value in a way that is seamless to the customer. Earlier in the series she introduced us to her Grandma (courtesy of our AI artwork tool). In this post on our Onchain Wallet Creation Tool, see if you can spot the part where she had to edit out swearing at herself for messing up the take 🤬 😁 On a serious note, as the tokenisation of real world assets (both tangible and intangible) explodes, many companies are going to be selling digital assets (NFTs etc.) to their customers. As we pointed out in Monday's post (see link in the comments), the native user experience associated with onchain wallets (where these assets will be managed and secured) is currently too confusing for the average user. That's why we've created tools like the one that Katie is demo-ing. We can seamlessly create an onchain wallet for a customer at any point in a user journey (e.g. at sign-up), and allow customers to access these wallets (and their contents) through the standard web or app experience (e.g. through their Account page via single sign-on). The customer doesn't even need to know that they are dealing with onchain assets. ⛓ However, if they want to full custody their digital assets, the tool allows them to transfer them to another wallet without having to worry about gas fees etc. Simple! ⛽ #crypto #blockchain #technology

5

8 Comments

The biggest reason for churn off AI products is that they lack accuracy or reliability. You can fix that problem for RAG applications in one line of code with Pongo's semantic filter. On a benchmark of 100 questions from Airbnb's 2023 10-k, we improved top 3 accuracy by 40% over OpenAI embedding vector search. Try 500 queries free -> https://joinpongo.com/ Benchmark results -> https://lnkd.in/gaStUZgq

18

4 Comments

Substrate is live! After nearly a year of development, really honing the developer experience and performance tuning a very unique runtime environment, we're finally releasing what we feel is by far the best way to run multi-inference workloads (e.g. agents, multimodal text, json, image, audio, embedding pipelines, etc). Substrate lets you use a ergonomic SDK (we really obsessed over the abstractions and typeaheads) to describe precise and complex relationships between AI tasks. And—importantly—it's connected to a backend specifically designed to run these graph workloads as efficiently as possible (optimized models, auto batching, insanely fast inter-node RPCs). Given the current state of "AI" models, we really believe that this pattern is the most effective for building capable, reliable, legible, extensible features with AI that go far beyond chatbots. We need both heuristic models like LLMs as well as normal symbolic computation, and a flexible/scalable way to coordinate these two modalities. We've already seen organizations like Substack building some super interesting pipelines that they can easily understand and iterate on, that just work at scale. It's still early in a much more ambitious vision for us, but we think this V1 is not only ready but already a meaningful improvement to many implementations. Reflecting on the last year, I feel so much gratitude for the team that I'm working with: Ben Guo, Liam Griffiths, Rishabh Parikh, 💭 Kyle Pitzen, Chris S., (& Hanwen Wu soon!). With this launch, we're announcing an $8M Series Seed led by Lightspeed (Nnamdi Iregbulem+Guru Chahal) with participation from many others. Feeling a ton of appreciation for the folks that took a bet on us or have lent their time and advice: South Park Commons (Ruchi Sanghvi+Aditya Agarwal), Craft Ventures, Red Swan Ventures (Andy Dunn), Lorimer Ventures, Chris Best, Jairaj Sethi, iqram and the immigrant groove, Andrew Kortina, Guillermo Rauch, Immad Akhund, Sam Lessin, William Gaybrick, Gokul Rajaram, Jana Messerschmidt, Shreyas Doshi, Matthew Hartman, Mike Cohen, Sabrina Hahn, Anne Lee Skates, Shreyans Bhansali, Julian Connor, Namrata Patel + many many others. https://lnkd.in/eHq8KUkA

120

9 Comments

We're shipping a lot here at Schematic, take a look at what went live this month. https://lnkd.in/ejmxiyi2 Highlights:  💲 Stripe integration: Sync billing data from Stripe and map it to company profiles in Schematic 🔌 SDK update: Caching, event buffering, and offline support for Go, Python, and Node.js ✨ UX improvements: Feature usage sidebar and bulk entitlements adding 🪵 Audit log: View, search, and filter updates to a company's features access, troubleshoot issues, and track changes for compliance 👥 Team permissions: Invite teammates and specify what permissions they have within Schematic

29

Legit quotable and actionable advice: * Power of Threes - 3 mentors, 3 peers, 3 mentees. * If you want to go fast, go alone. If you want to go far, go together

7

API Days NYC in May 2023 -- a year ago -- was the first API conference that I attended after founding Speakeasy. After one year, what's changed and what's stayed the same? It's clear that people are increasingly recognising how important it is to invest in APIs. Sessions like Jeremy Glassenberg's on API Product Management and Max Sadlowski's on how Wise was able to overhaul and maintain their API docs illustrates this. I also heard from Michaela Halliwell about how forward-thinking companies like HCSS are devoting more and more time and resources to ensuring a great internal and external API integrations. It's exciting to see the increased focus and attention that APIs are getting. However, there's still a huge gap between where the vast majority of APIs are today (manually / poorly maintained API docs), and what is necessary to deliver on the promise of ease of integration, rapid innovation and low cost-to-serve. We're looking forward to helping companies make a dent here with what we're building at Speakeasy. #apis #devex #sdks

13

2 Comments

PSA: Stop using the original OpenAI CLIP weights. I see a lot of people using CLIP for multi-modal retrieval. What many people may not know is that there are 100’s of CLIP models now since OpenAI’s original release. These newer models offer significantly better performance across many benchmarks. For product search related tasks we recommend the following. + Fastest inference: ViT-B-32 [laion2b_s34b_b79k] + Best balanced: ViT-L-14 [laion2b_s32b_b82k] + Best-all round: xlm-roberta-large-ViT-H-14 [frozen_laion5b_s13b_b90k] You might notice these are not the ones that perform the best on the benchmarks. Why? Because the benchmark only test for very specific tasks which are often quite different from search and recommendations. These are ones we have found to be the best for product related search and retrieval tasks and generally very good first choices if you are unsure where to start (even for non-product related search). Here is a hugging face space with all the data: https://lnkd.in/gxu5UYkt Read more: https://lnkd.in/g5_D4ccs

292

22 Comments

Got an email "GCP's Apache Kafka service is now live" at noon. After dinner, spent half an hour confirming it works well with Timeplus. I don't think you can make the bootstrap server publicly accessible. Has to setup a bastion host or run your app within the same project/VPC. Setting up the SASL username/pwd was not easy. Follow the guide carefully https://lnkd.in/gcKEVh3g and you may need to disable the iam.disableServiceAccountKeyCreation Anyway, if your workload are in GCP already, having such Kafka service in GCP is nice way to burn your cloud credits. Yet another Kafka vendor, great.

33

2 Comments

Live from New York, it's billions of rows in the browser 😎 Internal tools should be data-centric, not built as front-ends grafted onto your data. This means: 1. Tools can be auto-generated from data rather than configured in fiddly visual IDEs 2. Performance doesn’t suck when you pass 10k rows 3. You unlock agents that actually know how to operate your tools - not rely on weird browser RPA Had a blast presenting what Howard Zuo and I have been building at Dataland at the inaugural South Park Commons NY Demo Faire. Check our demo here: https://lnkd.in/e5wNrNjA Check the other incredible live demos here: https://lnkd.in/eHWp8U_S. All were great - my favorite was Speaksage :)

79

3 Comments

At Speakeasy we love working with devex leaders like Michael Heap to continuously raise the bar for what is "best in class" 🔥 We've been super lucky to work with him on what makes a great developer experience for iac and terraform. Come hear him and Thomas Rooney talk through how you can leverage OpenAPI and your API supply chain to build a provider and unlock the Terraform ecosystem. Link to the livestream in the comments 👇 #terraform #codegeneration #openapi

36

3 Comments

AI is just tech. It’s not magic and it doesn’t take a wizard with a PHD to understand how to utilize it. If you have setup and managed a database backed application in production before, then you’re 75% of the way there. If you’ve implemented Elasticsearch or another similar indexing service in the past, then you’re 95% of the way there. These skills are very transferable. While AI does still have a lot of growing up to do, now is the time to ramp up on it.

9