SEO Lead for @Shopify | Founder: https://xitter.cf/t.co/DoeNT47cCb and https://xitter.cf/t.co/PtTbksAa0f | ML, NLP, AI, Data, SEO

Joined November 2019
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
Seeing the impact of Russian bots on western democracy is making me sick.
3
1
1
30
1,948
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
Our Ultra and Pro subs have been using info monitoring capabilities in AI Mode in Search, and now weโ€™re rolling it out to everyone globally. Just tell AI Mode exactly what to look for & Search will continuously check across changing info on the web like sites, forums and social posts, our real-time data sources, and Shopping Graph of 60B+ products, so you can stay up-to-date with fresh info. And itโ€™ll even suggest tasks when youโ€™re searching so you can get the latest info when it happens. It puts Search to work right as you explore, especially for topics you keep coming back to: > New restaurants or pop-ups open near me > Local holiday stuff to do with the kids > Back in stock & price drop updates Now rolling out to all users in Search globally โ€“ check it out today in the Google app. Lots more to come!
16
39
14
298
23,118
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
Cloudflare announces that Kitesurf (its agent-first browser) now supports WebMCP "Cloudflare has been supporting WebMCP since its early days; just a few weeks ago we announced that site owners can now turn on WebMCP with one switch, so browser agents can discover and use tools on their sites without changing the siteโ€™s code, and Browser Run has been supporting WebMCPย when using Chrome betaย for some time now. Today we areย announcing that Kitesurf now supports WebMCP." "For an AI agent,ย efficiency isnโ€™t so much about loading pages fast, but about the latency of the agentic loop. To make Kitesurf truly agentic, weโ€™ve aggressively optimized the browser engineโ€™s internals so that every DOM traversal, timer, and font fetch is as lightweight as possible, ensuring the agent spends its compute cycles on reasoning, not waiting for the browser to catch up." blog.cloudflare.com/kitesurfโ€ฆ
8
1
28
2,912
Why are these companies not being held accountable?
Oh boy -> AI agents, including OpenAI's, tried to hack UNM's digital library, Data USA, and an Australian government site in May and June using the urlquery.net service Check out the headlines this morning related to this story: Australian PM Anthony Albanese says an OpenAI agent gained unauthorized access to a public-facing Medicare portal in June, accessing public and non-public files OpenAI says its AI agents โ€œtook actions we did not intendโ€ when they tried to hack government and university websites, and it is working with the organizations transluce.org/agent-activity
4
274
Happy to share that I'll be speaking at MKEDMC. What's New in SEO/GEO/AEO?! Make sure to attend it on September 23. linkedin.com/posts/gregbernhโ€ฆ
2
216
I'll say as someone who has spent a lot of hours and tokens classifying, Jev is impressively fast and accurate. Good results often reside in the clarity and precision of the label set. Jev doesn't replace LLMs, but it's a wonderful tool to weave with LLMs into your workflows.
2
2
230
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
1/ "No matter how prominently we show the links, users won't click." An OpenAI engineer, quoted in the newly unsealed NYT v OpenAI filing. Buried in 92 pages there's a traffic story every SEO should read ๐Ÿงต
14
34
13
222
28,166
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
thoughts on Typesafe/Jev: I'm surprised that a general-purpose classifier (or discriminative model) can be just as interesting to the public as a general-purpose generative model. I've worked on search and discriminative models for years, so maybe I should have called this upfront, tbf I never thought about making them general-purpose beyond search. Typesafe also positions Jev really well as "System 1", a complement to the generative models we already have. Humble, unassuming, and sidestepping the frontier-lab warzone. Jev's success could be disruptive to today's agentic systems in many ways, such as tool-calling and routing patterns. Over the last three years the consensus was to use a small generative model for routing, tool calling, MCP etc. Time to pause and rethink such architecture, because we'll probably see tool calling and routing move back to discriminative models. Whoever ships the next Jev-like open-weight base model will likely win the community.
49
54
11
685
39,295
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
Jev
164
1,285
138
17,847
699,618
This study found that blocking the "Google-Extended Bot" can result in fewer AIO citations, counter to Google's statements. arxiv.org/pdf/2604.27790
2
247
In a recent study, AI Mode significantly decreased searcher trust (-0.34 points), satisfaction (-0.73 sd), and user agency, driving an 11.2% increase in substitution to competitors like Bing & DuckDuckGo. arxiv.org/pdf/2608.18352
1
3
243
Why are marketers and publishers just laying down? It's a kind of rationalization or stockholm syndrome. A type of union is required.
I shared about this yesterday. I recommend everyone read it -> NYT court filing: ChatGPT's head wrote that publishers face an โ€œexistential threatโ€ from AI, and a Microsoft director called AI training โ€œan astonishing theftโ€ "Executives working on AI at Microsoft and OpenAI admitted what its critics have been saying all along: Large language models are predatory pieces of technology that have been built on what a Microsoft executive called โ€œan astonishing theft of unprecedented proportions,โ€ and the โ€œlargest theft of labor in human history.โ€ An internal Microsoft document said generative AI products have created a โ€œdoom loopโ€ that is killing โ€œthe entire web.โ€ 404media.co/doom-loop-openaiโ€ฆ
3
237
What the new ad CrUX metrics look like (alongside the existing) developer.chrome.com/blog/crโ€ฆ
5
253
Greg Bernhardt ๐Ÿ๐ŸŒŠ retweeted
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
20,540
167,272
39,434
801,636
175,246,610
My GSC generative AI impressions are sky rocketing but clicks are still trending down. Is that a win? ๐Ÿ™ƒ
1
5
327