Never trust your AI, unless you know where its brain is.
  • 21 Sep 2024
  • 4 Minutes to read
  • Dark
    Light

Never trust your AI, unless you know where its brain is.

  • Dark
    Light

Article summary

Hello, data aficionados and tech enthusiasts! Grab your favorite beverage, because I'm about to spill the tea (Earl Grey, naturally) on my recent escapade at the Big Data London show. Despite wrestling with a stubborn cold, I dove headfirst into this data-driven wonderland. Buckle up – it's going to be a wild ride!

The Great AI Illusion

Picture this: You're navigating a labyrinth of booths, each one screaming "AI" in neon lights. Exciting, right? Well, hold onto your bits and bytes, because here's the kicker – when I sat down with these companies, the conversation often went like this:

Me: "So, where's the AI in your product?" Tech guru, shuffling feet: "Well... actually... there isn't any."

It seems AI is the new black in marketing, but when it comes to actual implementation? It's more like the emperor's new clothes!

The OpenAI Bandwagon and a Wizarding Warning

Now, some companies were indeed using AI, but here's the twist: it was almost always OpenAI, bolted on like a spoiler on a family sedan. This got me thinking about a certain wise wizard's advice:

"Never trust anything that can think for itself, if you can't see where it keeps its brain."

Sage words, Mr. Weasley. In fact, I'd like to propose a new tech maxim:

"Never trust your AI, unless you know where its brain is."

What if OpenAI goes bust, or changes it's terms of service (again). Where might your data go?

With private models like SCOTi AI you know where the brain is and it safe with you on-prem.

Cybersecurity: The Plot Thickens

Before diving into the Big Data show, I attended the Cyber Security in Financial Services Summit. While I can't divulge all the secrets (some were more classified than a dragon's nest), here are two golden nuggets:

  1. Know your pipeline and dependencies like the back of your wand.

  2. Always have a Plan B (or C, or D) – because in the cyber world, constant vigilance is key!

Remember the MOVEit and CrowdStrike sagas? Billions of galleons – er, dollars – down the drain, and some companies didn't even know they were using these tools!

The Hidden Gems: Startup Edition

Now, let's talk about the fun stuff – the small, scrappy startups with big dreams. These are the booths where you can geek out with the founders themselves. Two caught my eye:

  1. Streambased: Imagine if SQL and  Kafka had a love child – that's Streambased. They've cracked some  impressive technical hurdles for speed and replication, staying true to  the spirit of Kafka. If you're into IoT, threat detection, or real-time  analytics, you might want to give these folks a shout. They're making  waves in the real-time data stream world!

  2. QuestDB:  An open-source time series database that's as lightweight as a Nimbus  2000. Their small footprint makes them perfect for specific  applications, and being open-source means you can take it for a test  flight with minimal risk. If you're dealing with timestamped data and  need efficient storage and analysis, QuestDB might just be your new best  friend.

The AI Unicorn in the Room

While most AI offerings were as bland as unseasoned porridge, one company stood out: Lemon AI. They're tackling the thorny issue of producing synthetic balanced datasets for training LLMs. At smartR AI, we've seen firsthand the challenges of insufficient or biased data when fine-tuning LLMs. While there are some concerns about model collapse with AI-generated data, Lemon AI's approach is intriguing enough that we're considering a trial project. Watch this space!

Old Friends and New Horizons

No tech show is complete without catching up with industry pals. I had a delightful chat with the Facts & Dimensions crew. Their impressive medical dataset is expanding to new frontiers as part of Filipe's quest to have all the world's public data in one place. I don't want to steal their thunder, but keep your ears open for some exciting announcements coming soon!

Other database darlings were also making waves:

  • Actian: Their "database for all  seasons" approach was drawing crowds. I had a great conversation with  Emma, their head of marketing, who mentioned they were harvesting some  promising leads from the show. Their variety of database engines seems  to be hitting the right notes with potential customers.

  • SingleStore:  These folks were busy as bees! Their platform's highly optimized  read/write performance is clearly striking a chord with the data-hungry  masses.

  • ClickHouse: Another booth buzzing with  activity. Their open-source, column-oriented OLAP database is resonating  well with customers craving real-time data solutions. If you need to  analyze data at lightning speed, ClickHouse might be your new best  friend.

The AI Conundrum

While AI was the buzzword du jour, it's clear that many exhibitors are still grappling with how to meaningfully incorporate it into their products. It's like everyone's trying to catch the Golden Snitch, but some are still figuring out how to mount their brooms!

For those of you curious about real AI solutions (ahem, SCOTi), don't be shy – reach out if we didn't connect in London. I'd be thrilled to chat about how we're pushing the boundaries of AI while keeping it private and under your control.

The Verdict

All in all, Big Data London was a data-driven delight, even if the "AI" part was more smoke and mirrors than substance. The database world is clearly alive and kicking, with innovative solutions popping up like Whack-a-Mole. Whether you're into streaming data, time series analysis, or just good old-fashioned data crunching, there's something for everyone in this ever-evolving landscape.

Until next time, keep your data big, your AI genuine, and your curiosity insatiable!


Oliver King-Smith is the CEO of smartR AI. Let me know what questions you want to ask SCOTi. Please feel free to send questions to oliverks@smartr.ai


Was this article helpful?