On-Prem

Storage

Digital memories are disappearing and not even AI or Google can help

Technology allows us to keep more of our stuff than previously possible – but what use is it if we can't find it?


Column I have too many artefacts detailing my digital history, which is stored on too many TRS-80 cassettes that hold sometimes-able-to-load BASIC programs, 3.5″ floppy disks, ZIP disks and HDDs.

All those storage devices are crowded with the rolling backup of my life over the last 20-ish years. Before that, my digital history sits in a twilight – some bits preserved, others lost forever.

Thirty years ago this month I sat down with Tony Parisi to write the first VRML browser, using Reality Lab before Microsoft re-badged it as Direct3D. I found that executable last year (which, props to Microsoft's focus on backward compatibility, still runs on Windows 11). But the source code has vanished completely – lost to some head crash back around 1996.

After enough such losses I became very conservative and careful. Everything I use is backed up – locally and to the cloud. The contents of computers I owned two decades ago reside as tiny folders in my multi-terabyte drives. I hardly ever lose anything anymore, which is wonderful – or would be, if I could find any of it.

Google wants us to search for everything – just dump it into a pile and let the search engine sort it all out. That sounds good in theory, but in practice it does not scale to the particular, and personal, and enduring. Not because the search engine fails – but because our human memories, crowded with experience, forget the exact language, or the precise dates, or any of the other criteria which might return an accurate hit.

Unlike the public web, our private data hasn't disappeared beneath a rising tide of AI-generated, SEO-optimized noise. We rely on memory and associative links that are not preserved and therefore cannot be located by any search. Google is simply not human enough to do a good job with our memories.

My lack of effective digital memory has become a serious issue. The amount of information I want to keep track of has grown unmanageably over the last twelve months. The explosive rise in research and applications of artificial intelligence has seen the number of open tabs in my browsers across six computers and mobile devices multiply from a handful into the hundreds. It's an affront to my tidy sense of digital order, a significant drag on browser performance – and a complete necessity.

All of those tabs are important. I just can't really remember what's open anywhere anymore – only that it's open. Somewhere.

So here I am, presented with a growing bill for twenty-plus years of personal technical debt. Perhaps I should have adopted a system of data cataloging beyond an endless number of folders in Dropbox and Google Drive. But what sort of system would have preserved the associative nature of all of our experiences?

Father-of-the VAX Gordon Bell's MyLifeBits – a project out of Microsoft Research, now 20 years old – considered the capture of experience, but nothing so far has provided an interface to the metadata of that experience. We need more than just time and location and association – it's context and continuity and meaning and mindset, and so much more that we can only barely articulate. Those nearly invisible threads bind us to our memories – both in our heads and in our digital selves.

So where does that leave me? I could magic up a GPT to do some basic analysis and metadata generation for the links I find in my research, the documents I read, and so on – but then what? How do these discrete items of digital memory maintain their context and continuity? How do I thread them together?

That it's hard to answer this question with any existing tool or technique tells us that we've found a blind spot. What we can't see, we can't move toward. In that darkness, more of our lives-as-context-driven-experience tumble down the memory hole – preserved but continuously fading into meaninglessness.

One of humanity's original technologies – the art of memory – urgently needs a reframing within modern technologies, allowing us to preserve the meaning of our personal past as we move into the future. ®

Send us news
71 Comments

Don't be fooled: Google faked its Gemini AI voice demo

PLUS: The AI companies that will use AMD's latest GPUs, and more

Google launches Gemini AI systems, claims it's beating OpenAI and others - mostly

Gemini accepts text, images, audio, and video and comes in three flavors

Google unveils TPU v5p pods to accelerate AI training

Need a lot of compute? How does 8,960 TPUs sound?

Google teases AlphaCode 2 – a code-generating AI revamped with Gemini

Don't worry, your developer jobs are safe … for now

Tech world forms AI Alliance to promote open and responsible AI

Everyone from Linux Foundation to NASA and Intel ... but some big names in AI are MIA

Creating a single AI-generated image needs as much power as charging your smartphone

PLUS: Microsoft to invest £2.5B in UK datacenters to power AI, and more

Google releases fix for missing Drive for desktop files

Just install the latest client and follow the instructions, but don't ask questions

Time to take action: Google's inactive account purge begins Friday

You should've received an email if you're affected, but here's a reminder just in case

Trust us, says EU, our AI Act will make AI trustworthy by banning the nasty ones

Big Tech plays the 'this might hurt innovation' card for rules that bar predictive policing, workplace emotion assessments

Google's Project Ellman: Merging photo and search data to create digital twin chatbot

'This is a brainstorming concept a team is at the early stages of exploring'

AI agents can copy humans to get closer to artificial general intelligence, DeepMind finds

Google’s AI offshoot finds copy-cat robots capable of aping living mentors

AI offers some novel crystal materials that could form future chips, batteries, more

What's more, a robot managed to cook some of them up. So, y'know, it might not be entirely science fiction