This week has been jam packed with traveling, meetings, events and all sorts! For an introvert like me, it's been pretty hard going pretending to be extroverted and interacting with lots of folks.
The biggest news this week was that my company won another award. A few weeks ago in September we won the CogX award for Best Fintech Company 2023. On Thursday I attended an awards ceremony with my colleague in order to accept the award for<ahref="https://thebusinessmagazine.co.uk/business_events/south-coast-tech-awards-2023/"> Best Emerging Tech Company (on the south coast) 2023</a>.
Doing my very best Chandler Bing photo smile face whilst holding our award with my colleague Alex (left).
</figcaption></figure>
<hr/>Earlier in the week I did a lot of traveling. On Monday I was up in our London office for a bunch of meetings and face to face catch ups with some new members of the team. Then, on Monday night I made my way over to a hotel near the London Excel centre and prepared for the <ahref="https://www.snowflake.com/events/data-cloud-world-tour-london/">Snowflake World Tour</a> on Tuesday.
<figcaption>The Snowflake World Tour was pretty glitzty and very high energy. Some of the content gave me real "Pied Piper" / "Silicon Valley" vibes</figcaption></figure>
We're not using Snowflake that heavily at work so it was an interesting opportunity for me to learn a little bit more about other features they offer and what we might want to start looking into. The new streaming features that allow you to build views that combine "streamed" data with "static" data looked pretty cool. One theme that kept coming up again and again was Large Language Models and Generative AI. However, very few of the speakers actually said what it is they thought you should be doing with these tools, merely that the platform "now supports them". I wrote in my notes that
<blockquote>
Every time llms come up it is always very vague and hand wavey</blockquote>
There was one demo which was actually pretty cool where they describe a database schema to a model and ask it to generate SQL queries based on natural language questions and then execute these queries and display the answers. This is actually a pretty reasonable use case for this tech because:
<ul>
<li>the risks of inconspicuous hallucination are low the (query either works or doesn't work)</li>
<li>if the SQL is invalid we can feed the error message back to the model and attempt to fix the query).</li>
<li>Sure the user can attempt to jailbreak the model or do prompt injection but we can use RBAC controls on the model's database access to make it read only and limit it's access to tables that we're happy the end user can see.</li>
</ul>
I had a play with <ahref="https://github.com/Snowflake-Labs/sfguide-frosty-llm-chatbot-on-streamlit-snowflake/blob/main/src/prompts.py">the demo </a>that the salesforce team shared afterwards and I was able to get it working with a local MySQL database (as opposed to a Snowflake instance) pretty damn quickly which was cool
<hr/>I stayed on in London on Tuesday night and on Wednesday morning I took a train down to Chelmsford to give a talk to some young professionals attending a networking event by the University of Essex as part of their KTP programme. I really enjoy public speaking these days so I had a great time and afterwards I got lunch with the University team and we chatted TV shows.
<hr/>A pattern I tried out this week, particularly at the conference on Tuesday was using my <ahref="https://github.com/usememos/memos">memos</a> instance in combination with <ahref="https://github.com/mudkipme/MoeMemosAndroid/">MoeMemos</a> a bit like twitter to "live tweet" interesting observations from the conference. This was a really nice way to do low friction note taking while enjoying the conference and adding photos is pretty seamless. I've got the <ahref="https://github.com/EINDEX/logseq-memos-sync">LogSeq Memos Sync</a> plugin which pulls in all the memos that I make and attaches them to my daily journal so that I can process and summarise my notes later. The other really cool thing about this is that tags in memos just transparently sync with tags in Logseq so if I tag something #snowflake in memos it will show up as a mention on my snowflake page in Logseq later on. Very groovy.
<hr/>This week I've not had a whole lot of recreational time and when I have had a bit of time I've felt like just snoozing or watching TV. I did watch the latest episode of Great British Bake Off on tuesday and Mrs R and I have started watching the latest series of Only Murders In The Building on Disney+.
On Saturday I went with Mrs R to visit my sister-in-law and we started watching Fall of the House of Usher on Netflix. This series is apparently loosely based on the works of Edgar Allan Poe and it's by Mike Flanagan who directed a bunch of shows I've previously enjoyed including <ahref="https://www.imdb.com/title/tt6763664/">Haunting of Hill House</a>, <ahref="https://www.imdb.com/title/tt10970552/">Haunting of Bly Manor</a> and <ahref="https://www.imdb.com/title/tt10574558/">Midnight Mass</a>. There is a lot I enjoyed about House of Usher except there is a particularly graphic/gory scene in one of the early episodes which is still etched permanently into my memory and left me feeling pretty panicked and prevented me from sleeping. The last thing that impacted me so deeply was <ahref="https://www.imdb.com/title/tt0914798/">The Boy in Striped Pajamas</a>. I don't know if I want to keep watching the show as whilstt I'm a big fan of psychological horror, I don't do gore and normally Flanagan excels in the former rather than the later. However, I'm kind of captivated by the storyline so I'm wondering if I should just keep pushing through and hope that it won't cause any more lost sleep.
<hr/>I've not read any more of Ringworld by Larry Niven this week, I'm kind of put off by the in-your-face misogyny at this point. I may DNF it and move on to something else.
I did pick up <ahref="https://www.amazon.co.uk/Think-Like-CTO-Alan-Williamson/dp/1617298859">Think Like a CTO </a>(which is probably a good idea since I am one) this week which I'm enjoying so far.
<hr/>This week should be a lot less hectic and busy. We are going to the theatre to see 2:22 on Tuesday, I'm only in London for one day this week for some meetings on Wednesday and I'm taking a half day on Friday to visit my parents and some friends up in the midlands whilst Mrs R stays at home and celebrates her sister's birthday with a girly spa afternoon.
