AI is what’s making both the above a real possibility
This is why we’re building Knowatoa to try to help people navigate what is clearly going to be a much more fractured and complicated search future.
PerplexiTok
Perplexity AI is entering the TikTok acquisition ring, and honestly, it’s a bit like watching your nephew who just learned Javascript, declare they could rebuild NASA’s mission control software over a weekend.
Their pitch?
Rebuild TikTok’s “black box” recommendation algorithm from scratch
Make the “For You” feed open source
Add “Community Notes features” similar to what X (formerly Twitter) has
Perplexity’s positioning as a middle ground (claiming they have “world-class technical capabilities” but with “Little Tech independence”) misses the point that the TikTok ban is politically motivated, and nobody cares about the company’s technical capabilities.
It’s hard to see this going through with the head of a different social search AI company so actively involved in the White House.
Adam Mosseri, head of Instagram, said during his weekly AMA that Instagram is working on expanding IG’s search to work better for content and not just creators.
I don’t know if we’ll ever see Meta.ai become a straightforward competitor to Google (or even ChatGPT).
But we will see the underlying LLama AI that Meta has invested billions of dollars in become the underpinning of massive search features in Facebook, Instagram, and WhatsApp.
Social search is often a huge blind spot for SEOs as it doesn’t have the same visibility as traditional search, but billions of dollars of commercial searches are conducted every year on social platforms.
This week, the awesome Anisah Osman Britton interviewed me for Sifted about creating an SEO strategy for AI search.
Like all of Sifted’s content, they tried to keep things very practical and actionable, and I was thrilled to see that this was true for all of the AI and search questions they asked.
I hope other startup founders reading it were inspired to at least start laying the groundwork for their AI search strategy.
I’d encourage you to read the article, but if you just want to dive in, you should read the BISCUIT framework, as it’s a little more formal in its structure on how to make this happen.
One person’s “malicious and out of control” AI bot scraping attack is another person’s “near real-time AI indexing” and into that Labyrinth we go.
Quite literally, Cloudflare has released a new tool that shunts “misbehaving” bots to an AI, generating a Labyrinth of content to waste the scraper/indexer’s resources.
While I’m all for site operators protecting their sites from downtime, this seems incredibly risky to me as it’s literally (yes, I used that word again, yes, I’m using that work correctly, don’t worry) spewing out weird AI-spun content on your domain. You just have to trust that Cloudflare didn’t mess it up and actually serve this content to a bot you care about now or in the future.
Note: This is exactly why we built the “AI Search Console” feature in Knowatoa: so you can at least test and monitor whether the well-known bots can crawl your site. It’s available on all plans (even Free), so you should check it out.
Last week, I mentioned that Claude was getting a new Web Search feature, and Simon Willison did some digging into both Anthropic’s legal Terms and Services and their code to discover that they’re using Brave’s search index under the hood.
Brave is more well known for its focused browser than its search index, but if you’re interested in adding your own site to the index, the instructions are on their community forum: How to add website to Brave Search
I’ve mentioned them before, but I’m continuing to really enjoy hanging out in the SEO Community. It’s a Slack group of four thousand really kind and helpful people.
