top of page

Monday, July 22, 2024

Photo of man with arms crossed
Dave Findlay CEO Fuse Data

This week's edition


Happy Monday all!


When it comes to conversations in the data space, some weeks are hot and some weeks are not.


Last week was certainly hot. My feed was full of amazing posts and conversations!

Picking the top three proved to be a real challenge this week, but here they are.


In this issue, we'll be diving into:


📈 What's hot and what's not: Gartner releases their hype cycle for data management


🤖 The AI hype train continues to slow down


💡 What data teams can learn about from Obama


😄 Your Weekly Dose of Data Hilarity.


I hope you enjoy the issue!


Best regards,

Dave


📈 What's hot and what's not: Gartner releases their hype cycle for data management


Consultant Gregor Zeiler provides a great recap in his post of the major changes in the Gartner Hype Cycle for Data Management this year.


Items that I found most interesting were points made around the data lakehouse architecture and the data mesh framework.


Zeiler notes that, as per Gartner, the data lakehouse architecture will reach a plateau before data lake. This makes sense to me, given the balance that lakehouse strikes between speed, structure, and governance. I would even go a bit further to say that pure play data lakes will be considered obsolete before they reach the plateau.


The other interesting point in the post is that Gartner is declaring data mesh obsolete. While I believe the full vision of data mesh is likely difficult to achieve (even books on the topic use fictitious companies vs real-world examples), I still like aspects of it. I find myself going back to data mesh principals often, but adapting them to the realities of my client's culture and day-to-day operations. So, in short, while it's hard to implement in its purest form, I believe there is a ton of good in the framework. Just pick what works for you and don't feel like it has to be an all or nothing situation.


🤖 The AI hype train continues to slow down


Author and speaker Joe Reis offers up a post with a refreshing message for those that are getting tired of the GenAI hype.


In the post and the accompanying Substack article, Reis expresses frustration with the constant vendor push to sell AI into organisations without any real value present. Reis cautions buyers about previous data hype cycles around the big data and data science movements that left organisation with lots of tooling, but little in terms of value being delivered.


I agree with everything Reis puts forward here and even pointed in this direction in an earlier issue of Data Clarity. Every vendor you come across now has an AI story, but it's up to us as data professionals to get beyond the story, understand the value, and decide whether or not we even need AI to solve the problem at hand.


💡 What data teams can learn about from Obama


The biggest problem I hear when talking to CIOs is that data teams are just not getting things done. They worry about how the business perceives their team and are frustrated at the progress being made. I can understand this, after all, talking about solving a problem is much easier than rolling up our sleeves and solving it, right? 🤪


Vendors will have you believe that if you just pick the right architecture, framework, or technology, you will be able to meet the needs of your organisation. But that's not the case. I've often joked with customers that I could build an enterprise data platform on MS Access. I'm glad nobody has taken me up on the task, but the point is made to illustrate that the tooling selection isn't as important as it seems. So don't over-rotate on that stuff.


Actually meeting the needs of your organisation involves getting down to work and getting things done. It means meeting with stakeholders to understand needs, having tough prioritisation discussions around fixed capacities and commitments, it involves putting egos aside and listening to critical design feedback, shipping to production, and working with stakeholders on change management to ensure adoption is there. This is the hard work, but it's what drives results.


So when I saw this video from Barack Obama, posted by entrepreneur Steven Bartlette, on the importance of getting stuff done, I had to include it in this newsletter.


😄 Your Weekly Dose of Data Hilarity


As long as someone is on call to hit refresh!


ree

Want this delivered to your inbox?





Monday, July 8, 2024

Photo of man with arms crossed
Dave Findlay CEO Fuse Data

This week's edition


Happy Monday all!


As the temperatures continue to skyrocket across the country so do the data conversations that I've been watching!


In this issue, we'll be diving into:


🤖 What do Citadel and Nintendo have in common?


💰 What to consider, besides total cost, when evaluating new data tools.


🦹‍♂️ Shot's fired! Are the data heroes at your organization really villains in disguise?


😄 Your Weekly Dose of Data Hilarity.


I hope you enjoy the issue!


Best regards,

Dave


🤖 What do Citadel and Nintendo have in common?


It's really hard to think of any similarities between the US hedge fund and the Japanese game and device maker, however, there is one - they are both a bit cool on AI. While most companies are riding the AI hype train, executives at Citadel and Nintendo seem to be a bit more measured.


This post from technology consultant Dr. Jeffery Funk highlights excerpts from a Fortune article where Citadel founder and CEO Ken Griffin describes the current limits of LLM and other AI models and believes the talent that Citadel is hiring is better equipped to deal with the realities of the business world. Griffin notes that models do well when there is consistency in the underlying data, however, regimes change, and the real world is full of inconsistencies. Something that human intuition is much better equipped to deal with today.


Nintendo's take is similar in some ways, as they are also betting on their own talent and don't really see value in leveraging generative AI for game making. A post by Amir Satvat highlights Nintendo President Shuntaro Furukawa's thoughts on the matter:


We have decades of know-how in creating optimal gaming experiences for our customers, and while we remain flexible in responding to technological developments, we hope to continue to deliver value that is unique to us and cannot be achieved through technology alone.

These two executives are certainly pouring a bit of cold water on all the AI hype, but it is welcome in my opinion. Bad (and costly) decisions are made when technological advances are riding the hype curve. Level-headed executives need to be able to get ahead of this to protect budgets and time investments from their teams and companies.


All of that, however, may be coming to an end. The most recent Gartner hype cycle for AI suggests that the hype on generative AI has peaked and is descending into the trough of disillusionment. This is not a bad thing at all. It's in this trough that true value is built. This is where companies can really dig in, build ROI positive solutions, and help pull this innovation out of the trough and into the plateau of productivity.


💰 What to consider, besides total cost, when evaluating new data tools.


When evaluating new tools, most companies will only look at the sticker price when developing their business case, but there are many other factors that should be considered as well.


That's why I was happy to see a post last week from Saks VP of Data, Veronika Durgin, that laid out a comprehensive list of factors depending on your decision to either "build" or "buy" the tools.


The only item I would add to Veronika's list would be the cultural fit of the selection. This is, in some ways, covered by the training and onboarding; however, it's a slightly different flavour. Admittedly, it's a bit quantitative in nature, but it can have a significant impact on the cost of getting things done.


For example, if you're a digital native enterprise with a strong engineering culture, you may not find a fit by implementing a low-code platform like Informatica or Talend as your data integration and management solution. Even after training, the technology philosophy may not be accepted by your team... and nobody wants a grumpy data engineering team. 🤣


🦹‍♂️ Shot's fired! Are the data heroes at your organization really villains in disguise?


Last week, Nicholas Mann, CEO of Stratus Consulting, decided to take on the entire data engineering community in his post that leads off with:

Data engineers who insist on custom coding everything only care about one thing. Themselves. 

The point that Nicholas is trying to make is that leaders need to be mindful of team members who may be believers in the old "job security through obscurity" philosophy.


Cowboy coding practices and the like can, on the surface, appear to solve problems, but in reality, they create technical debt and single points of failure. Nicholas's post is a very good reminder to be vigilant for this in your organization, even if he did take some heat in the comments.


In my experience, I've seen a few data villains, but most of the "cowboy coding" was done with the best intentions. Usually it comes as a result of 1) underinvestment in the data program and 2) pressure from leadership to "just get the data". These two factors, when put to a resourceful individual, will only lead to one outcome. Proper investment in your data program and skilled leadership will most certainly mitigate the issues that Nicholas has put forward.


😄 Your Weekly Dose of Data Hilarity


.... let me Google that for you.

ree

Brought to you by Consulting Comedy.




Want this delivered to your inbox?





Monday, July 1, 2024

Photo of man with arms crossed
Dave Findlay CEO Fuse Data

This week's edition


Happy Canada Day all! 🇨🇦


This week marks the re-launch of our newsletter!


Dose of Data Clarity will now come out on Monday mornings (ET) and it will focus on synthesising the hot data topics and discussions that are going on in the industry.


In this issue, we'll be diving into:


🤖 LLMs are amazing, but are they being applied to the right problems?


🌎 Dark data and the environmental impact of forgetting how to:

drop table d_customer_master_dev_test2;

 

🥊 OpenAI, LLaMA or Claude? The Salesforce LLM rankings are in.


😄 Your Weekly Dose of Data Hilarity


I hope you enjoy the issue!


Best regards,

Dave


🤖 LLMs are amazing, but are they being applied to the right problems?


Last week's AP article by Wyatte Grantham-Philips, which discussed the difficulties faced by McDonald's AI-driven drive-thru, generated quite a bit of discussion in the data community.


The IBM-powered solution, while impressive in many regards, fell short of its required accuracy. According to Cate Chapman from LinkedIn News, 1-in-5 orders required human intervention. It's clear that having 20% of drive-thru customers experience friction in their ordering process is not something that was acceptable to McDonalds, as the program with IBM is coming to an end this month at all test locations.


Even though the experiment didn't work for McDonalds, Grantham-Philips notes that other brands have successfully implemented similar technology. Popeyes even goes as far as to say their solution is 97% accurate at capturing orders, however, some people may have a different take on that figure. 🤣


Stories like these really underscore that understanding user needs and the delivered user experience is paramount. Without that, will anyone actually use the stuff we build? 😅


🌎 Dark data, AI and the environment


Author and data executive Sol Rashidi generated some conversation on LinkedIn with her thoughts on dark data and the energy consumption of LLMs like ChatGPT.


According to Sol, ChatGPT consumes half a litre of water for every 5-10 prompts, and an AI generated image consumes enough power to charge your smartphone to 24%! Another post suggests that a query can consume the equivalent of running a 10-watt LED light bulb for 18 minutes to 3 hours. That's quite the energy footprint!


The flip side of this is the topic of dark data. This is data that resides in your environment and is not being used for any purposes. It's collecting virtual dust. It's using up space on your servers, which equates to energy, and if this data is hooked up to pipelines its footprint is larger, especially when you consider the overhead of maintaining things that aren't being used.


Data Strategist Malcolm Hawker provides his take on the issue of dark data in his Forbes article. Malcolm offers up some good suggestions on dealing with dark data, which is becoming increasingly important these days as more companies are setting ambitious ESG goals.


I guess now is as good a time as any to take a look in our own (server) closet, figure out what's being used and what isn't, and do a little digital spring cleaning! 🌱


🥊 OpenAI, LLaMA or Claude? The Salesforce LLM rankings are in.


Last week, Salesforce AI Chief Clara Shih posted the company's LLM benchmark for CRM. Shih notes that this is an ever evolving space, with new models dropping all the time, the foundational model layer becoming increasingly commoditised.


This is absolutely correct. A quick check on Hugging Face shows over 700,000 models equipped to handle pretty much any task that you can think of.


With rapid commoditisation in the model layer, Shih points out that customers should consider using an offering (like the Salesforce AI stack) that acts as an abstraction layer and allows interchangeability between the off-the-shelf model of the day and BYO models.


The results of the Salesforce study also point out that out-of-the-box model accuracy is just one dimension when selecting your models. Enterprise users also need to consider cost and speed when making their choice.


When you put it all together, it looks like Claude is becoming a real contender! 🥊


😄 Your Weekly Dose of Data Hilarity

ree




We've all been here at one point in our careers.... right?


Brought to you by Joe Fenti, via Instagram.






Want this delivered to your inbox?





fuse data logo
bottom of page