going bananas

We had some more bananas go black, so time for some more banana oat buns.  Yum.  Unlike last time, they didn't try to escape.

superpowers complete!

I got my annual NHS flu jab about three weeks ago, but am not eligible for an NHS Covid jab this year.  So I decided to pay for one.  £98.95.  Ouch!  That hurts more than the needle!

Having booked online at a pharmacy (since my GP doesn't offer them privately), I went to get the jab today.  So, I am now invulnerable again.  Well, not for 7 days, for it to take effect.  And then, of course, not 100% guaranteed.  But still, waaaay better than the alternative.

word of the day: slopsquatting

I have grave doubts about LLMs and their ilk: there is no "there" there.  But that will have to wait for a future blog post.  One thing they are known for is "hallucinating", or generating things that have plausible form, but no denotation, such as citing non-existent legal cases, scientific references, and ... code modules.

This last one is new to me, discovered in a comment on Ask a Manager.  I write my own code, of course.  But apparently some people are happy to have AI write code for them.  I suppose yet another form of AI slop shouldn't surprise me.  However, the associated security vulnerability is even more worrying than some random bogus citations.  It's the problem of slopsquatting:

Like other forms of gen AI, coding AI makes up references to non-existent code libraries. “Slopsquatting” is when a malicious actor creates malicious code with the name of one of these non-existent code libraries. When you run your AI-generated code, instead of throwing an error it automatically downloads and runs that malicious code.

This is why we can't have nice things.

daylight savings costing coding

We have four sources of data we combine to get our various solar power usage charts.  We have data from the production of the solar panels on the house (PV1&2) and separate data from the new solar panels on the garage roof (PV3).  We have data from a Wattson meter on how much power we consume in the house, and data from the Zappi on how much we use to charge the car (either overnight off the grid, or during the day from excess solar).  Combining these is tricky, because they all have different granularities (from every 30s to every hour).

The combination is even trickier when the clocks change.  None of the data sources tell us whether the timestamps are in GMT, BST, or something else, but we have guessed, probably correctly.  When the clocks went back last week, the Wattson and PV3 systems simply repeated the times for the relevant hour, with the different data.  But the PV1&2 and Zappi systems overwrote the first hour of data.  Aaargh.

Thankfully for my sanity at least, it's my other half who combines the data.  I use the combined data to plot the charts.  So he was tearing his hair out, rather than me doing so.  He has now completed the task.

But still.  Honestly, it's not as if daylight savings is a new thing!  It reminds me of the very first item on the list of Falsehoods programmers believe about time.  At least it gives me an excuse to re-read the list, anyway.)

still retired

A year ago today was the first day of my retirement.  As I mentioned at the time, I have several academic retirement projects on the go.  So how has the first year gone?

Pretty well, actually. I have been to three conferences: CapoCaccia in May, UCNC in August, and ALife in October.  I have had multiple papers published, some of which I did a large chunk of the work for. I have started learning about Topological Data Analysis, and am applying it to some real biological data sets; I've really enjoyed writing all the code for this.  Our new LoCoMo ARIA project is fascinating, with our two post-docs making great strides, and I am pulling ideas from both simulation and open-ended evolution into it.  My penultimate PhD student has passed his viva.  Added to this, I have a handful of other papers and projects also making progress.  

And I think I have figured out how to level up the open-ended evolution experiments I have been thinking about, thanks to a chance conversation at ALife that put me on the beginning of the track to a relevant 2003 paper (many thanks to whoever that was; I spoke to so many people there I'm afraid I have forgotten who pointed me in this direction originally).  But it's still just pages of scribbled notes and fever dreams for now; watch this space!

On the non-academic side, I am managing to do more reading, too.

So, all in all, a productive, enjoyable, and actually amazingly relaxing, first year.  Not having a fortnightly 175-mile commute is a wonderful lifestyle change.  It did take me about 6-9 months to stop going: "wait, what is it I'm forgetting to do?"  Particularly when doing some of that fiction reading.  But I'm now past that, and looking forward to year two of fun research.