Keeping (mostly) up to date with the rapidly changing world of search and digital marketing so you don’t have to.
Thursday 9th December
December product review update bigger than April
While the overall article is summed up by the above title, there’s an interesting and in-depth dissection of ‘exclusive data’ from RankRanger and Semrush that details exactly what made the update bigger and possible consequences that is well worth a read on the Search Engine Land blog.
Wednesday 8th December
I’m pretty sure there’ll be a lot of articles on the topic this coming year, but this one from Digiday caters to some of my own particular biases – including the following quote from Sébastien Borget, COO of the metaverse platform The Sandbox:
“Accessibility is at the heart of main-stream adoption and this is why sometimes the simplest is often the ‘go-to’ path, enabling anyone with just an internet connection and a computer to enter the metaverse will be a winner. With the evolution of VR and AR and potential upcoming new devices touted by the top tech companies, these could become new enjoyable ways of accessing it too.”
The summary is that there is currently a disconnect between what is being promised by Zuckerberg, what is actually achievable in the short-term and what the end user understands the metaverse to be. It will be interesting to see how it shakes out.
New CWV Metric
Search Engine Journal features a story on the upcoming introduction (potentially) of a new core web vitals metric. You can read more on the subject and why we think it will be introduced next year here.
Inclusive language at Google
With inclusivity one of the trends we look at in our upcoming predictions eBook, it was interesting to see Google’s Mueller and Splitt hosted podcast tackle the topic of language inclusivity. You can find a write-up on Search Engine Roundtable.
Tuesday 7th December
Image extensions for desktop
Search Engine Land is featuring a piece on the upcoming rollout of image extensions for desktop. Advertisers using them on mobile will have them automatically added to appropriate desktop ads in a change that will take a couple of weeks to fully apply.
Monday 6th December
TikTok gifts and tips
This may be a real concern to other social media platforms – TikTok is set to allow creators to directly receive monetary rewards via the platform. The platform has been the fatest to achieve a billion monthly active users, almost halving the time it took Instagram, and this may prove a powerful incentive for creators to add TikTok to their strategy (think YouTube creators summarising longer videos, allowing them to diversify the source of their income).
Thursday 2nd December
Image optimisation plugins
Not really news, but I’ve been meaning to check out some options on this for a while, so I assume it’s on the list for other people too. SE Journal has some reviews on their blog.
Wednesday 1st December
Product Reviews Update December 2021
Set to roll out over a three week period, the update is intended to prioritise reviews with a focus and expertise within a subject domain – meaning that affiliate sites with huge numbers of product reviews crossing ranges, types and technologies may lose out to single subject review sites. Search Engine Journal covers the update in more detail here.
Google Pathways AI
I’ve been spending most of my spare time looking through Open AI documents to see whether I can get my head around it and which projects it might help with – it’s incredible. If you’d told me when I was trying to code in QBasic back in the early 90s that computers would be able to do this stuff by 2021, I would have been at least sceptical if not entirely dismissive.
10 PRINT “HA HA NO IT WONT”
However, if a recent blog from Googler Jeff Dean is taken at face value, Google could well be taking the next steps with their Pathways AI. For anyone new to the idea, your general machine learning model will be trained on one data set or type of data set (whether that’s photo-recognition or the Atari game ‘Video Pinball’) and will master that specific data set. However, in order to be useful in processing another data set it needs to be retrained from the ground up. Deep Mind was able to self train to beat Atari games as early as 2015, but it took until 2017 before it was able to transfer some learning between Atari games in order to master multiple games without starting from scratch.
It then took a further 2 years before an ML model was able to do that with any kind of speed. The ambition of Pathways AI is to allow the use of a single ML model for millions of tasks – and for it to be able to transfer learning from any of those tasks to develop unique approaches to a range of situations – much as a mammal might.
…we’d like to train one model that can not only handle many separate tasks, but also draw upon and combine its existing skills to learn new tasks faster and more effectively.
That way what a model learns by training on one task – say, learning how aerial images can predict the elevation of a landscape – could help it learn another task — say, predicting how flood waters will flow through that terrain.
We want a model to have different capabilities that can be called upon as needed, and stitched together to perform new, more complex tasks – a bit closer to the way the mammalian brain generalizes across tasks…
The blog deals with the future tense at the moment, but getting to this stage suggests a reasonable level of progress (at least to the extent where feasibility must have been established). This is quite an exciting development for ML and AI and, while it’s likely at least a couple of years off, it could be revolutionary not just for computing but for the world in general.
Want to keep up with the latest search and digital marketing news? Why not subscribe to our blog, or contact us to see what we can do to help keep your brad ahead of the curve.