Calvin Fu
@calvinfu1128.bsky.social
I sit at the intersection of UX, strategy, data, and digital product.
I’m not building hands-on, but I’m always shaping what gets built — and why.
I’m not building hands-on, but I’m always shaping what gets built — and why.
“Optimizing for AI search” sounds great, until you try to do it.
Each AI tool uses different logic. No real ranking signals to influence.
Agencies can help polish content, but visibility in AI answers? That mostly comes down to your product, your credibility, and what people really say about you.
Each AI tool uses different logic. No real ranking signals to influence.
Agencies can help polish content, but visibility in AI answers? That mostly comes down to your product, your credibility, and what people really say about you.
Google AI, ChatGPT rarely agree on brand recommendations: Data
Google's AI Overviews and AI Mode and ChatGPT disagree on brand picks nearly two-thirds of the time, new data shows.
searchengineland.com
September 3, 2025 at 12:33 PM
“Optimizing for AI search” sounds great, until you try to do it.
Each AI tool uses different logic. No real ranking signals to influence.
Agencies can help polish content, but visibility in AI answers? That mostly comes down to your product, your credibility, and what people really say about you.
Each AI tool uses different logic. No real ranking signals to influence.
Agencies can help polish content, but visibility in AI answers? That mostly comes down to your product, your credibility, and what people really say about you.
I use AI browsers to preview articles & help me decide what’s worth a full read.
I already avoid anything sensitive or login-based. But Brave’s discovery about Comet’s prompt injection risks...That still shocked me a bit.
These tools are promising. But they’re not ready for blind trust.
I already avoid anything sensitive or login-based. But Brave’s discovery about Comet’s prompt injection risks...That still shocked me a bit.
These tools are promising. But they’re not ready for blind trust.
Perplexity Comet Browser Vulnerable To Prompt Injection Exploit
Vulnerability discovered in Perplexity's Comet AI browser enables attackers to gain access to sensitive data in open browser tabs
www.searchenginejournal.com
September 1, 2025 at 7:57 AM
I use AI browsers to preview articles & help me decide what’s worth a full read.
I already avoid anything sensitive or login-based. But Brave’s discovery about Comet’s prompt injection risks...That still shocked me a bit.
These tools are promising. But they’re not ready for blind trust.
I already avoid anything sensitive or login-based. But Brave’s discovery about Comet’s prompt injection risks...That still shocked me a bit.
These tools are promising. But they’re not ready for blind trust.
Funny (or worrying?): ChatGPT said OpenAI’s browser is “Comet.”
Comet? That’s Perplexity’s.
OpenAI’s is Operator.
When AI hallucinates its own company’s product, we really need to ask: how much can we trust these tools at face value?
Comet? That’s Perplexity’s.
OpenAI’s is Operator.
When AI hallucinates its own company’s product, we really need to ask: how much can we trust these tools at face value?
August 27, 2025 at 11:52 PM
Funny (or worrying?): ChatGPT said OpenAI’s browser is “Comet.”
Comet? That’s Perplexity’s.
OpenAI’s is Operator.
When AI hallucinates its own company’s product, we really need to ask: how much can we trust these tools at face value?
Comet? That’s Perplexity’s.
OpenAI’s is Operator.
When AI hallucinates its own company’s product, we really need to ask: how much can we trust these tools at face value?
Yes, users want memory in AI. But not like this.
When context slips away mid-conversation, or memory cuts off sessions—with no warning—that’s not intelligent design.
We need memory that works, not just marketing memory.
When context slips away mid-conversation, or memory cuts off sessions—with no warning—that’s not intelligent design.
We need memory that works, not just marketing memory.
Sam Altman on GPT-6: 'People want memory'
OpenAI CEO Sam Altman says the next version of ChatGPT will be more personal than its predecssors.
www.cnbc.com
August 21, 2025 at 4:18 AM
Yes, users want memory in AI. But not like this.
When context slips away mid-conversation, or memory cuts off sessions—with no warning—that’s not intelligent design.
We need memory that works, not just marketing memory.
When context slips away mid-conversation, or memory cuts off sessions—with no warning—that’s not intelligent design.
We need memory that works, not just marketing memory.
Casual users may be fine with one-size-fits-all LLMs.
But power users? We’ve hit the limits: sluggish tools, hallucinations, vague outputs.
SLMs are now emerging—faster, cheaper, domain-specific.
Will this finally give us clarity… or just create a new maze of choices?
But power users? We’ve hit the limits: sluggish tools, hallucinations, vague outputs.
SLMs are now emerging—faster, cheaper, domain-specific.
Will this finally give us clarity… or just create a new maze of choices?
NVIDIA Says Small Language Models Are the Future of Agentic AI
INVEST IN US BY AUGUST 28 TO EARN UP TO 25% BONUS SHARES. Welcome to Atlas.
www.linkedin.com
August 21, 2025 at 2:29 AM
Casual users may be fine with one-size-fits-all LLMs.
But power users? We’ve hit the limits: sluggish tools, hallucinations, vague outputs.
SLMs are now emerging—faster, cheaper, domain-specific.
Will this finally give us clarity… or just create a new maze of choices?
But power users? We’ve hit the limits: sluggish tools, hallucinations, vague outputs.
SLMs are now emerging—faster, cheaper, domain-specific.
Will this finally give us clarity… or just create a new maze of choices?
When companies promote "AI-powered" everything, we need to ask: is it real innovation—or branding wrapped in jargon?
GEO vs SEO is just one example.
We don’t need new acronyms. We need better products.
AI should empower, not just advertise.
GEO vs SEO is just one example.
We don’t need new acronyms. We need better products.
AI should empower, not just advertise.
Google Suggests AI SEO Acronyms May Lead To Scams & Spam
Google's John Mueller replied to a lot of the hype around all GEO, AIO, AEO, etc topics floating around the SEO space. John said on Bluesky, "The higher the urgency, and the stronger the push of new ...
www.seroundtable.com
August 16, 2025 at 2:35 AM
When companies promote "AI-powered" everything, we need to ask: is it real innovation—or branding wrapped in jargon?
GEO vs SEO is just one example.
We don’t need new acronyms. We need better products.
AI should empower, not just advertise.
GEO vs SEO is just one example.
We don’t need new acronyms. We need better products.
AI should empower, not just advertise.
Google’s Preferred Sources feature is live for Top Stories.
Some fear echo chambers. But here’s the truth:
Diversity is a user choice.
• Stick to big names? You’ll see them.
• Read niche voices? They’ll rise.
Personalization = reflection.
Some fear echo chambers. But here’s the truth:
Diversity is a user choice.
• Stick to big names? You’ll see them.
• Read niche voices? They’ll rise.
Personalization = reflection.
Google Top Stories Preferred Sources Graduates Labs
Google has launched the Preferred Sources top stories feature fully in the US and India, after testing it in Search Labs for a bit over six weeks. As I said then, this is a nice feature for publishers...
www.seroundtable.com
August 14, 2025 at 12:36 PM
Google’s Preferred Sources feature is live for Top Stories.
Some fear echo chambers. But here’s the truth:
Diversity is a user choice.
• Stick to big names? You’ll see them.
• Read niche voices? They’ll rise.
Personalization = reflection.
Some fear echo chambers. But here’s the truth:
Diversity is a user choice.
• Stick to big names? You’ll see them.
• Read niche voices? They’ll rise.
Personalization = reflection.
Not every user needs deep model control.
But as a ChatGPT Plus user, GPT‑5’s auto-routing felt too opaque — it crashed my browser, slowed everything down.
But as a ChatGPT Plus user, GPT‑5’s auto-routing felt too opaque — it crashed my browser, slowed everything down.
ChatGPT's model picker is back, and it's complicated | TechCrunch
GPT-5 was supposed to get rid of all the complicated AI models to choose from in ChatGPT, but that's not how it's panned out.
techcrunch.com
August 14, 2025 at 1:18 AM
Not every user needs deep model control.
But as a ChatGPT Plus user, GPT‑5’s auto-routing felt too opaque — it crashed my browser, slowed everything down.
But as a ChatGPT Plus user, GPT‑5’s auto-routing felt too opaque — it crashed my browser, slowed everything down.
Perplexity offered to buy Chrome.
Symbolic - but not random either.
They’re building the same control stack: browser, engine, monetization, defaults.
Sound familiar?
Are we witnessing disruption—or just the birth of the next Google?
Symbolic - but not random either.
They’re building the same control stack: browser, engine, monetization, defaults.
Sound familiar?
Are we witnessing disruption—or just the birth of the next Google?
Perplexity offers to buy Chrome for billions more than it's raised | TechCrunch
Perplexity tells TechCrunch the terms of the offer include a commitment to keep Chrome’s underlying engine, Chromium, open source and to invest $3 billion into it.
techcrunch.com
August 13, 2025 at 1:52 AM
Perplexity offered to buy Chrome.
Symbolic - but not random either.
They’re building the same control stack: browser, engine, monetization, defaults.
Sound familiar?
Are we witnessing disruption—or just the birth of the next Google?
Symbolic - but not random either.
They’re building the same control stack: browser, engine, monetization, defaults.
Sound familiar?
Are we witnessing disruption—or just the birth of the next Google?
Elon Musk wants to inject ads directly into Grok.
It’s a logical next step—AI infra and talent aren’t cheap, and advertisers are eager to follow attention.
But will users accept paid suggestions inside conversations?
It’s a logical next step—AI infra and talent aren’t cheap, and advertisers are eager to follow attention.
But will users accept paid suggestions inside conversations?
Elon Musk to introduce ads to X’s AI chatbot
Billionaire owner outlines plans to improve advertising on social media platform in rare overture to marketers
www.ft.com
August 12, 2025 at 1:27 AM
Elon Musk wants to inject ads directly into Grok.
It’s a logical next step—AI infra and talent aren’t cheap, and advertisers are eager to follow attention.
But will users accept paid suggestions inside conversations?
It’s a logical next step—AI infra and talent aren’t cheap, and advertisers are eager to follow attention.
But will users accept paid suggestions inside conversations?
Auto-routing glitches and browser instability in GPT-5 have made my post-launch experience bumpy
Sam Altman addresses ‘bumpy’ GPT-5 rollout, bringing 4o back, and the ‘chart crime’ | TechCrunch
The Reddit AMA got spicy as users peppered OpenAI with questions, and some asked the company to bring back its previous model.
techcrunch.com
August 9, 2025 at 10:47 PM
Auto-routing glitches and browser instability in GPT-5 have made my post-launch experience bumpy
When GPT-5 lost track of hours of work that had been working together…☠️
Still trying to get the work back…what a nightmare
Still trying to get the work back…what a nightmare
August 9, 2025 at 5:09 AM
When GPT-5 lost track of hours of work that had been working together…☠️
Still trying to get the work back…what a nightmare
Still trying to get the work back…what a nightmare
Browser crashed several times already using GPT-5 today, anyone experiencing it too?
August 9, 2025 at 3:21 AM
Browser crashed several times already using GPT-5 today, anyone experiencing it too?
Auto-routing in GPT-5 means less friction, but also less visibility. Would you want to choose? Personally, I want
OpenAI's GPT-5 is here | TechCrunch
OpenAI CEO Sam Altman says GPT-5 is the "best model in the world," and aims to make ChatGPT more intuitive to use.
techcrunch.com
August 8, 2025 at 2:52 AM
Auto-routing in GPT-5 means less friction, but also less visibility. Would you want to choose? Personally, I want
GPT‑5 now turns prompts into live software agents. Useful? Absolutely. Dangerous if unchecked? Possibly.
OpenAI launches GPT-5, nano, mini and Pro — not AGI, but capable of generating ‘software-on-demand’
With safer design, more robust reasoning, expanded developer tooling, and broad user access, GPT-5 reflects a maturing AI ecosystem.
venturebeat.com
August 7, 2025 at 10:57 PM
GPT‑5 now turns prompts into live software agents. Useful? Absolutely. Dangerous if unchecked? Possibly.
AI Overviews might not be killing traffic, says Google. Instead, they claim it’s making traffic more meaningful. What do you think—spin or insight?
Google swears it isn’t destroying the web with AI search
Changes to Search have sent ripples across the media industry.
www.theverge.com
August 7, 2025 at 11:33 AM
AI Overviews might not be killing traffic, says Google. Instead, they claim it’s making traffic more meaningful. What do you think—spin or insight?
GitHub’s Thomas Dohmke reimagines developers as ‘creative directors of code,’ urging adoption of AI tools to enhance system design and QA. A call for evolution, not replacement.
GitHub CEO delivers stark message to developers: Embrace AI or get out.
Thomas Dohmke wrote that humans are often resistant to change. He said that's okay, but these people should probably find another profession.
www.businessinsider.com
August 7, 2025 at 10:12 AM
GitHub’s Thomas Dohmke reimagines developers as ‘creative directors of code,’ urging adoption of AI tools to enhance system design and QA. A call for evolution, not replacement.
Anthropic's Claude has quietly become the #1 choice for businesses (32% market share), overtaking OpenAI's 25%.
What's really interesting: this happened in just 2 years.
Makes you wonder - what are enterprises seeing in Claude that consumers aren't prioritizing yet?
techcrunch.com/2025/07/31/e...
What's really interesting: this happened in just 2 years.
Makes you wonder - what are enterprises seeing in Claude that consumers aren't prioritizing yet?
techcrunch.com/2025/07/31/e...
Enterprises prefer Anthropic's AI models over anyone else's, including OpenAI's | TechCrunch
Anthropic holds 32% of enterprise LLM market share by usage. This is a sharp reversal from just two years ago when OpenAI held 50%.
techcrunch.com
August 1, 2025 at 4:46 AM
Anthropic's Claude has quietly become the #1 choice for businesses (32% market share), overtaking OpenAI's 25%.
What's really interesting: this happened in just 2 years.
Makes you wonder - what are enterprises seeing in Claude that consumers aren't prioritizing yet?
techcrunch.com/2025/07/31/e...
What's really interesting: this happened in just 2 years.
Makes you wonder - what are enterprises seeing in Claude that consumers aren't prioritizing yet?
techcrunch.com/2025/07/31/e...
OpenAI is crawling llms.txt while Google isn't planning to. Should the AI player come up with a standard about it?
#OpenAI #Google #GEO #AEO
www.seroundtable.com/openai-crawl...
#OpenAI #Google #GEO #AEO
www.seroundtable.com/openai-crawl...
OpenAI Crawling LLMs.txt Files? Google Says It Won't.
Recently, Google said that no AI system is currently using the LLMS.txt file. But maybe some are starting to? OpenAI may be starting to discover and crawl LLMS.txt files on websites. While Google's G...
www.seroundtable.com
July 28, 2025 at 3:34 AM
OpenAI is crawling llms.txt while Google isn't planning to. Should the AI player come up with a standard about it?
#OpenAI #Google #GEO #AEO
www.seroundtable.com/openai-crawl...
#OpenAI #Google #GEO #AEO
www.seroundtable.com/openai-crawl...
In the past or even now, if you don't rank well in organic search, you may still be able to gain visibility through paid search
It seems it may not be the case in the AI era. Keep an eye on how the placements of paid advertising evolve when AI becomes more dominant
#AI
koddi.com/blog/ai-hote...
It seems it may not be the case in the AI era. Keep an eye on how the placements of paid advertising evolve when AI becomes more dominant
#AI
koddi.com/blog/ai-hote...
Metasearch meets AI: Hotel listings now live in Gemini - Koddi
Discover how Google’s Gemini AI is transforming travel search with embedded metasearch hotel ads throughout its AI results.
koddi.com
July 18, 2025 at 8:24 AM
In the past or even now, if you don't rank well in organic search, you may still be able to gain visibility through paid search
It seems it may not be the case in the AI era. Keep an eye on how the placements of paid advertising evolve when AI becomes more dominant
#AI
koddi.com/blog/ai-hote...
It seems it may not be the case in the AI era. Keep an eye on how the placements of paid advertising evolve when AI becomes more dominant
#AI
koddi.com/blog/ai-hote...
Not easy but I do hope they could bring disruption to the market, especially to OTAs
#LLM #AI #DirectBooker #OTA #disruption
www.hospitality.today/article/dire...
#LLM #AI #DirectBooker #OTA #disruption
www.hospitality.today/article/dire...
DirectBooker aims to give hotels a direct line into ChatGPT
www.hospitality.today
July 18, 2025 at 8:06 AM
Not easy but I do hope they could bring disruption to the market, especially to OTAs
#LLM #AI #DirectBooker #OTA #disruption
www.hospitality.today/article/dire...
#LLM #AI #DirectBooker #OTA #disruption
www.hospitality.today/article/dire...
It may disrupt the browser market. However, also have to be cautious about user data usage and privacy
#OpenAI #Google #browser #disruption #privacy #data
www.techradar.com/computing/fo...
#OpenAI #Google #browser #disruption #privacy #data
www.techradar.com/computing/fo...
OpenAI’s rumored AI web browser could feast on your data, new report says – and I'm worried for my privacy
A new kid on the browser block
www.techradar.com
July 11, 2025 at 5:25 AM
It may disrupt the browser market. However, also have to be cautious about user data usage and privacy
#OpenAI #Google #browser #disruption #privacy #data
www.techradar.com/computing/fo...
#OpenAI #Google #browser #disruption #privacy #data
www.techradar.com/computing/fo...
Most importantly is the content brings value to users
#Google #AI #Search #SEO #SearchEngineOptimization
ahrefs.com/blog/ai-gene...
#Google #AI #Search #SEO #SearchEngineOptimization
ahrefs.com/blog/ai-gene...
AI-Generated Content Does Not Hurt Your Google Rankings (600,000 Pages Analyzed)
Does Google punish AI-generated content? We analyzed 600,000 pages to find out.
ahrefs.com
July 9, 2025 at 2:56 AM
Most importantly is the content brings value to users
#Google #AI #Search #SEO #SearchEngineOptimization
ahrefs.com/blog/ai-gene...
#Google #AI #Search #SEO #SearchEngineOptimization
ahrefs.com/blog/ai-gene...
Then the question is...should links / citations be this dynamic?
#Google #AIMode
www.seroundtable.com/google-ai-mo...
#Google #AIMode
www.seroundtable.com/google-ai-mo...
Google AI Mode Desktop vs Mobile & Same Browser
It seems that the links/citations you see in Google AI Mode can be very different when you search on desktop versus mobile. In fact, it seems like the sources can differ even if you try the same quer...
www.seroundtable.com
July 9, 2025 at 2:46 AM
Then the question is...should links / citations be this dynamic?
#Google #AIMode
www.seroundtable.com/google-ai-mo...
#Google #AIMode
www.seroundtable.com/google-ai-mo...