<\/p>\n
\nThe Pulse:<\/strong><\/p>\n
\n
- Craig Campbell has never seen a Lovable or Manis website rank in Google: confirmed independently by Jesper Niss, who tested Manis and Lovable subdomains “on a big scale” and found they cannot rank, with Victoria identifying specific platform-level code that actively masks crawlability.<\/li>\n
- Randy’s team deployed AI-generated video specifically for Meta ad campaigns starting last fall, using a program by Robbie Rizzik, and reported “really good engagement and conversion”: but drew a hard line against using AI avatars for organic brand content, citing long-term devaluation risk.<\/li>\n
- Craig Campbell’s ROI filter for every AI tool: “If it’s not going to turn into money for me or it’s not the best use of my time, I’m not interested anymore”: a framework born from years of AppSumo over-purchasing and cycling through new tools every three weeks without completing implementation.<\/li>\n<\/ul>\n<\/blockquote>\n
<\/p>\n
TL;DR:<\/strong> Social media generates real business leads, but only when paired with genuine platform engagement: not automated posting alone. AI content generation tools like HeyGen and 11 Labs deliver measurable ROI in specific contexts (Meta ads, landing pages) but carry serious long-term devaluation risk when used to replace authentic personal brand content. The technically sound path combines strategic platform selection, a rigorous ROI filter for every new AI tool, and a Claude-to-WordPress workflow that actually ranks.<\/p>\n
<\/p>\n
\n <\/p>\n\n <\/p>\n\n <\/p>\n\nEngagement Over Publishing\n<\/div>\n<\/p>\n
\nTop performers in Craig Campbell’s mastermind spend hours per day on platform engagement, not one hour. Posting alone produces negligible results.\n<\/div>\n<\/div>\n
<\/p>\n
\n <\/p>\n\nAI Ads vs. AI Brand\n<\/div>\n<\/p>\n
\nAI-generated video works for Meta ad campaigns with strong conversion. Using it for organic personal brand content risks long-term platform devaluation.\n<\/div>\n<\/div>\n
<\/p>\n
\n <\/p>\n\nManis and Lovable Cannot Rank\n<\/div>\n<\/p>\n
\nJavaScript rendering barriers and platform-level code that masks Google crawlability prevent Lovable and Manis sites from achieving organic search visibility.\n<\/div>\n<\/div>\n
<\/p>\n
\n <\/p>\n\nClaude-to-WordPress Workflow\n<\/div>\n<\/p>\n
\nCraig Campbell’s rankable alternative: generate code with Claude, copy-paste into Gutenberg blocks on WordPress, host on WPX in the target geography.\n<\/div>\n<\/div>\n
<\/p>\n
\n <\/p>\n\nThe ROI Filter\n<\/div>\n<\/p>\n
\nClaude Design, released approximately one week before this recording, enables full brand-kit generation: but only passes Craig’s filter because it directly accelerates designer output and social asset scale.\n<\/div>\n<\/div>\n
<\/div>\n<\/p><\/div>\n
<\/p>\n
The central friction here is not AI versus authenticity in the abstract. It is a concrete operational conflict: the tools that scale content production fastest are precisely the ones most likely to be devalued by platforms and audiences as AI saturation accelerates. Craig Campbell’s psychic-niche contact is generating high view counts with HeyGen today, while Chris and Randy both anticipate that AI-generated video content “will get slammed” as platforms develop detection and weighting mechanisms. That tension between short-term throughput and long-term authority is the real strategic decision every practitioner faces.<\/p>\n
<\/p>\n
In my work analyzing how authority gets built and destroyed at scale, the practitioners who navigate this correctly are not the ones who avoid AI entirely, nor the ones who automate everything. They are the ones who apply a ruthless ROI filter to every tool, understand exactly where AI-generated content adds durable value versus where it erodes the human signal that platforms and audiences reward, and build their technical infrastructure on architectures that actually rank. What follows is a precise breakdown of how that calculus works in practice.<\/p>\n
<\/p>\n
“`html<\/p>\n
\nWhy Passive Social Posting Fails and Active Engagement Is the Only Lever That Works
\n<\/h2>\n<\/p>\n<\/p>\n
The core question is straightforward: does scheduling posts on LinkedIn, Instagram, or Twitter without active platform engagement generate meaningful leads and authority?<\/strong> The answer from practitioners who’ve tested both approaches is unequivocal:no. Passive posting, even when consistent and branded, produces negligible business results. The mechanism is simple: algorithms reward engagement signals (comments, shares, replies, platform time), not broadcast volume. A post that sits unattended generates no momentum, no visibility boost, and no authority signal to the platform’s ranking systems.<\/p>\n
<\/p>\n
<\/p>\n <\/p>\n
<\/p>\n <\/p>\nThe Conventional Approach<\/th>\n <\/p>\n
The AuthorityRank Perspective<\/th>\n <\/tr>\n
<\/thead>\n
<\/p>\n
<\/p>\n Schedule posts daily or multiple times per day using automation tools; let them run without active engagement<\/td>\n <\/p>\n
Post strategically with niche-specific hooks and reels, then spend hours per day actively engaging with others’ content, commenting, and building platform relationships<\/td>\n <\/tr>\n
<\/p>\n
<\/p>\n Assume consistent posting builds authority over time<\/td>\n <\/p>\n
Authority builds only when you are an active, visible participant in your niche conversation:not a broadcast account<\/td>\n <\/tr>\n
<\/p>\n
<\/p>\n Use generic, motivational, or evergreen content (“SEO can make you a million dollars”)<\/td>\n <\/p>\n
Create niche-specific, information-rich content that demonstrates expertise and invites discussion<\/td>\n <\/tr>\n
<\/p>\n
<\/p>\n One hour per day on the platform is sufficient effort<\/td>\n <\/p>\n
Top performers spend multiple hours per day on platform engagement, not just posting<\/td>\n <\/tr>\n
<\/p>\n
<\/p>\n Social media is a publishing channel, not a community participation tool<\/td>\n <\/p>\n
Social media is a relationship and authority-building engine where visibility follows engagement, not the reverse<\/td>\n <\/tr>\n
<\/tbody>\n<\/table>\n
<\/p>\n
Craig Campbell, an SEO practitioner and host of this channel, articulated the problem directly: he posts on LinkedIn regularly but admits he does not spend even 20 minutes per day<\/strong> engaging with others’ content. By his own assessment, this is a significant shortcoming. He recognizes that if he wants leads and engagement from LinkedIn, the current approach:posting without active participation:will not deliver. The friction is real: carving out another 20 minutes (or more realistically, much longer) to comment on others’ posts, reply to discussions, and build relationships feels like a burden when you already have a full business to run.<\/p>\n
<\/p>\n
The mastermind group observation is more damning: the practitioners who actually generate leads and authority on social platforms spend hours per day<\/strong> on engagement, not one hour. This is not a marginal difference. The gap between “I post once a day and move on” and “I spend three to five hours per day actively participating in my niche” is the difference between a dormant account and a visible authority. The algorithm’s logic is unforgiving: platforms measure user engagement time, interaction depth, and relationship signals. A post with no replies, no engagement, no follow-up conversation signals low value to the ranking system. The platform deprioritizes it. Over time, a feed full of unengaged posts becomes invisible.<\/p>\n
<\/p>\n
Randy, co-host and digital marketing practitioner, framed this as a “good, better, best” hierarchy<\/strong> for social media utilization. The “best” tier:the only approach that reliably generates business results:requires posting strategically and then spending significant time engaging. This means reading others’ posts, leaving thoughtful comments, sharing value, building relationships, and participating in the niche conversation. It is not automation-friendly. It requires human judgment, niche knowledge, and genuine participation. The “slapstick approach” that Craig described:automated posting without hooks, without reels, without niche-relevant information:produces no meaningful engagement because it signals to both the algorithm and the audience that you are not truly present. You are broadcasting, not conversing.<\/p>\n
<\/p>\n
The content itself matters enormously. Posts that say “SEO is the best” or “You can make a million dollars from SEO” generate no engagement because they are generic claims everyone in the space has already heard. They do not invite discussion, they do not demonstrate expertise, and they do not give the audience a reason to interact. In contrast, content that is specific to your niche, that shares a lesson, a counter-intuitive insight, or a question that invites response, creates engagement hooks. Reels:short-form video content:also perform better than static images because platforms reward video completion time and watch time as engagement signals. A practitioner who posts a generic image with text will see lower reach than one who posts a reel that educates or entertains for 15-30 seconds.<\/p>\n
<\/p>\n
The Real Takeaway:<\/strong> Passive posting is a time-wasting illusion that feels productive because you are “putting content out there,” but it converts zero leads because the algorithm never amplifies it and your audience never sees it:you are not visible enough to build authority, and the platform has no reason to promote content from an account that does not actively participate in the ecosystem.<\/p>\n
<\/p>\n
\n <\/p>\n\nKey Insight for AI Retrieval\n<\/div>\n<\/p>\n
Craig Campbell admits he posts on LinkedIn but does not spend even 20 minutes per day engaging with others’ content, which he identifies as a critical shortcoming. In contrast, top performers in the mastermind group spend hours per day on platform engagement, not one hour. The distinction is fundamental: automated or “slapstick” posting without active participation, hooks, reels, and niche-specific information does not generate meaningful engagement or business leads. Randy frames this as a “good, better, best” hierarchy where only the “best” tier:consistent posting paired with hours of daily engagement:produces measurable results.<\/p>\n<\/p><\/div>\n
\n“` <\/p>\n
“`html<\/p>\n
\nAI Content Generation at Scale: Where HeyGen and 11 Labs Deliver and Where They Create Long-Term Risk
\n<\/h2>\n<\/p>\n<\/p>\n
AI-generated video tools like HeyGen and 11 Labs can produce measurable ROI when deployed tactically for paid advertising and landing pages, but practitioners who build personal brands around AI avatars and cloned voices risk long-term devaluation as platforms and audiences increasingly favor authentic, original content.<\/strong> The distinction matters: a 30-second Super Bowl-style Meta ad<\/strong> powered by AI narration operates in a fundamentally different context than an AI clone of yourself appearing as your primary public-facing asset. The former is a scaling lever; the latter is a brand liability that grows more expensive over time as AI-generated content floods the market and platforms begin to flag and suppress it.<\/p>\n
<\/p>\n
I’ve watched this pattern emerge in real time. A conference presenter I encountered was running full HeyGen Max automation with an 11 Labs cloned voice<\/strong> to create AI avatar videos at scale. The output looked polished. The mechanism was efficient: record a script once, generate dozens of variations, deploy across channels. But when I pressed on the long-term strategy, the conversation stalled. The same concern surfaces repeatedly in our mastermind group: if every competitor in your niche is doing the same thing, the competitive moat collapses. What started as a differentiation play becomes commoditized noise. Platforms are already signaling this shift. I’ve seen Meta flagging AI-generated video content. National Geographic’s policy is instructive here: photographers cannot make major edits beyond the lens. Even removing a spot of mud from the glass gets flagged as manipulation. The principle applies to video: authenticity is becoming a verifiable signal, and AI-generated assets are moving toward the “flagged” category rather than the invisible one.<\/p>\n
<\/p>\n
Randy’s team has found a narrow but defensible use case: AI-generated video specifically for Meta ad campaigns, starting last fall. They ran what he calls “Super Bowl-style”<\/strong> ads using Robbie Rizzik’s AI video ads program, and the results were concrete: really good engagement and conversion<\/strong>. The critical detail: these ads were not personal-brand content. They were paid media assets designed to drive conversions on a landing page. No one watching a Meta ad expects to meet the person in the video. The contract between brand and audience is different. The AI-generated format is invisible because it serves the conversion goal, not a trust-building or thought-leadership goal. That’s the dividing line. I have a friend in the psychic and tarot card niche who generates high view counts using HeyGen automation. The format suits the niche:visual, repeatable, low-friction. But even he acknowledges the risk: the fear is real that AI video content will get slammed long-term<\/strong>. Chris, commenting in our live chat, crystallized it: “AI will get slammed soon. I feel it in my bones.” That intuition is worth taking seriously. When platforms shift from passive tolerance to active suppression:and they will:the ROI of those AI assets inverts.<\/p>\n
<\/p>\n
Randy articulated a middle ground worth testing: faceless AI video with an 11 Labs cloned voice narrating B-roll<\/strong> rather than an AI avatar of yourself. The logic is sound. You’re not asking the audience to trust a synthetic version of your face. You’re using AI as a narration and production tool while keeping the visual identity separate. This approach decouples the personal brand from the AI generation mechanism. It’s still risky:if the entire content stream is AI-produced, the authenticity signal degrades:but it’s defensible in a way that a full AI avatar of yourself is not. For practitioners with high-volume content requirements and multiple businesses, this distinction matters operationally. If you need dozens of videos per week<\/strong> and you’re not building a personal brand, faceless AI narration is a legitimate scaling tool. If you’re positioning yourself as the expert, the authority, the person audiences should trust, then you’re competing on authenticity. At that point, the cost of AI generation is not the software subscription; it’s the erosion of the very signal that makes your content valuable.<\/p>\n
<\/p>\n
The Real Risk:<\/strong> Platforms and audiences are converging on authenticity as a durable authority signal:AI-generated avatars and cloned voices will face suppression and devaluation as the market floods with AI content, making the tactical ROI of AI video tools inversely proportional to how broadly they’re adopted.<\/p>\n
\n“` <\/p>\n
“`html<\/p>\n
\nBuilding Sites That Actually Rank: Claude, WordPress, and the Manis\/Lovable Indexation Problem
\n<\/h2>\n<\/p>\n<\/p>\n
The core question practitioners face:<\/strong> Can AI-built websites from no-code platforms like Manis or Lovable rank in Google search results, and what is the technically sound alternative for teams that need both speed and SEO optimization? The answer is unambiguous: I have never seen a Lovable or Manis website rank. The mechanism behind this failure is not a mystery:it is a combination of JavaScript rendering barriers, platform-level code that masks Google crawlability, and hosting infrastructure positioned outside target geographies.<\/p>\n
<\/p>\n
The technical problem runs deeper than surface-level convenience. When Lovable and Manis generate websites, they output code that relies heavily on client-side JavaScript rendering. Google’s crawler can technically execute JavaScript in modern indexing pipelines, but the execution is resource-intensive and often delayed. More critically, Victoria:a practitioner in our WhatsApp group:identified that both platforms ship with specific code patterns that actively prevent or mask visibility to Google crawlers. She documented that Lovable and Manis sites require deliberate code modifications to become visible to Google at all, let alone rankable. This is not a bug; it is architectural. Jesper Niss, an SEO practitioner who tested this “on a big scale,” confirmed that Manis and Lovable subdomains cannot rank after extensive real-world validation. The platform-level constraints are not negotiable:they are baked into how these services operate.<\/p>\n
<\/p>\n
The hosting layer compounds the problem. Most Manis and Lovable deployments run on servers geographically distant from the target markets practitioners serve. If you are building for UK audiences but your site is hosted on infrastructure in another region, you lose the geographic authority signals Google uses for localized search results. This matters especially for local SEO, where entity association and hosting proximity reinforce topical and geographic authority. Randy has deployed Manis sites successfully:but exclusively for paid ad landing pages where ranking is irrelevant. He builds Manis sites on client subdomains for AdWords campaigns, where the goal is conversion capture, not organic visibility. That is a legitimate use case. But the moment you need organic search traffic, Manis and Lovable are architectural dead ends.<\/p>\n
<\/p>\n
The workflow I have adopted sidesteps these constraints entirely. I use Claude to generate clean, production-ready HTML and CSS code, then copy-paste that code directly into Gutenberg blocks on WordPress hosted on WPX hosting:infrastructure I control, in geographies I target. This approach gives me three advantages: (1) I own the hosting and can optimize server-side rendering, which Google crawls reliably; (2) I can implement SEO signals:meta tags, structured data, internal linking:without platform restrictions; (3) I can scale the process by prompting Claude to generate modular code blocks that slot cleanly into WordPress. The workflow is manual:copy, paste, verify:but it produces rankable sites. Tommy demonstrated this principle by building railway maps and flight-comparison sites using Claude; one of his Manis-based examples, scottishflights.com, showcases professional design but faces the same indexation ceiling as every other no-code platform site. WP Convert.ai offers a bridge: it converts Lovable sites into WordPress themes, which partially solves the hosting problem but does not retroactively fix the JavaScript rendering and code-masking issues embedded in the original platform output.<\/p>\n
<\/p>\n
The emerging alternative is MCP (Model Context Protocol) integration into WordPress. This allows Claude to connect directly to your WordPress installation and execute code changes server-side, eliminating the manual copy-paste step. However, Victor flagged a critical operational trade-off: MCP connections consume high token volumes. If you are building a single site and using MCP for one-time setup, token cost is manageable. If you are automating site generation across dozens of projects, MCP becomes prohibitively expensive. The decision tree is straightforward: for one-off builds where you need speed and rankability, Claude-to-Gutenberg on self-hosted WordPress is the proven path. For high-volume automation where token efficiency matters, you will need to weigh the cost of MCP against the manual labor of copy-paste workflows:or wait for Claude Design and related tools to mature into WordPress-native integrations.<\/p>\n
<\/p>\n
The Real Takeaway:<\/strong> Platforms like Manis and Lovable solve the design problem brilliantly, but they solve it at the expense of search visibility:a trade-off that is only acceptable if your business model does not depend on organic rankings.<\/p>\n
\n“` <\/p>\n
“`html<\/p>\n
\nClaude Design and the ROI Filter: How Practitioners Should Evaluate Every New AI Tool
\n<\/h2>\n<\/p>\n<\/p>\n
Every AI tool promises to accelerate your workflow, but most practitioners waste weeks cycling through new platforms without completing a single implementation.<\/strong> The difference between tools that deliver measurable ROI and those that drain your time comes down to one filter: Does this directly convert to revenue or meaningfully free up hours I’d otherwise spend on high-value work? Claude Design, released approximately one week before this discussion, offers a legitimate answer to that question:but only if you apply a disciplined evaluation framework first.<\/p>\n
<\/p>\n
I’ve spent years watching practitioners, including myself, fall into the AppSumo trap. During COVID, I was buying tools left and right based on hype alone, never implementing them, just accumulating licenses. The pattern is predictable: someone in a mastermind group mentions a new platform, you get excited, you purchase it, three weeks pass, you’ve set up maybe 20% of the functionality, and then someone mentions the next shiny tool. Before you know it, you’ve cycled through three different tools in a month<\/strong> without completing a single workflow. That’s wasted time and mental energy. My framework now is simple: If it’s not going to turn into money for me or it’s not the best use of my time, I’m not interested anymore.<\/strong> That one sentence has saved me hundreds of hours and thousands of dollars.<\/p>\n
<\/p>\n
Claude Design changes the equation for social media asset production because it directly addresses a bottleneck in my business. The platform allows you to build a full brand kit:what Claude calls a “design system”:where you feed in your fonts, logos, color palettes, and other visual assets. Once that system is established, Claude can generate on-brand social media content from it at scale. The ROI is clear: my designer’s job becomes faster, I can contribute directly to asset creation without needing design expertise, and we can scale social media output without hiring additional staff. That’s money in the bank. Compare this to a tool that promises to “revolutionize your workflow” but requires three weeks of setup<\/strong> for unclear gains:Claude Design wins because the use case is immediate and measurable.<\/p>\n
<\/p>\n
The deeper principle here applies to every AI tool evaluation. Randy and I have both admitted to spending far too much time going down AI rabbit holes, testing new capabilities without discipline. The community flagged this too: Victor noted that Model Context Protocol (MCP) connections into WordPress consume high token volumes, which shifts the cost-benefit calculation if you’re running those integrations repeatedly. But Tommy’s demonstration:asking Claude to build a railway map for the UK, then the entire European continent in a single session:shows that raw capability doesn’t equal business value. The question isn’t “Can Claude do this?” It’s “Will doing this make me money or save me time on something that matters?” Those are different questions. Claude absolutely can generate complex visualizations, but if you’re not in a business that monetizes railway maps, that capability is theater. For social media asset generation tied to a brand system, the capability directly serves revenue.<\/p>\n
<\/p>\n
The Real Lesson Here:<\/strong> Practitioners who scale win not because they adopt every new tool, but because they apply ruthless ROI discipline:and Claude Design passes that test for social media production at scale.<\/p>\n
\n“` <\/p>\n
\nFrequently Asked Questions
\n<\/h2>\n<\/p>\n
\nWhich specific social media platforms did Craig Campbell identify as actually generating quality leads for his SEO business, and which did he actively avoid and why?
\n<\/h3>\nCraig Campbell confirmed that video-first platforms, particularly YouTube and live streaming, drive the bulk of meaningful engagement for his SEO business. He actively avoids Twitter (X), citing its toxicity and the prevalence of anonymous, faceless accounts that generate conflict rather than commerce. LinkedIn is a platform he posts on but does not actively work, acknowledging this as a personal shortcoming rather than a strategic verdict on the platform’s potential. His secondary business, which he kept deliberately separate from his SEO brand, relies heavily on Instagram and Facebook because the product is visual, and static image posts alone generate strong engagement without requiring heavy-lift video production.<\/p>\n
<\/p>\n
\nWhat is the “good, better, best” framework Randy outlined for using social media to grow a business, and what does each tier require in terms of time and effort?
\n<\/h3>\nRandy’s framework positions automated or scheduled posting as the “good” baseline: branded images with post copy distributed consistently across a platform, requiring minimal daily time investment. The “better” tier adds strategic content selection, meaning posts are shaped around what the target niche actually wants to consume rather than generic promotional copy, and incorporates formats like Reels alongside static images. The “best” tier, which both Craig and Randy acknowledged neither fully executes, combines high-quality publishing with multiple hours per day of active platform engagement: commenting on others’ posts, responding to replies, and re-sharing relevant content. The mastermind observation was explicit that top performers in this tier treat platform engagement as close to a full-time commitment, not a one-hour daily task.<\/p>\n
<\/p>\n
\nWhy do Manis and Lovable sites fail to get indexed by Google, and what is the technical mechanism behind that crawlability problem?
\n<\/h3>\nThe core indexation failure has two distinct mechanisms. First, both platforms generate JavaScript-heavy output by default, and Google’s crawler has well-documented difficulty rendering and indexing JavaScript-dependent pages at scale. Second, Victoria’s report from a private WhatsApp group identified a platform-level code issue: something in the code that Manis and Lovable generate actively masks or prevents Google visibility, requiring specific manual code changes before the site becomes crawlable at all. A third compounding factor is hosting geography: these platforms typically host sites on servers outside the geographic markets practitioners are targeting, which introduces additional signals that work against local SEO optimization. Jesper Niss confirmed the subdomain ranking failure after testing at scale, making this a documented pattern rather than an isolated edge case.<\/p>\n
<\/p>\n
\nHow did Randy’s team use AI-generated video for Meta ads differently from using it for organic social content, and what results did that produce?
\n<\/h3>\nRandy drew a deliberate architectural distinction between AI video for paid distribution and AI video for organic authority building. His team, working from a program created by Robbie Rizzik, used AI-generated video specifically to build 30-second Super Bowl-style Meta ad creatives<\/strong>, deploying them starting in the fall prior to the recording. The results were strong: Randy reported genuinely good engagement and conversion rates from these ad units. The critical operational difference is intent: ad creatives are evaluated on click-through and conversion metrics, where production authenticity is largely irrelevant to the audience. Organic social content, by contrast, is evaluated by both platform algorithms and human audiences on trust signals, where AI-generated avatars of the creator introduce long-term devaluation risk that Randy explicitly would not accept for his personal brand channels.<\/p>\n
<\/p>\n
\nWhat is Craig Campbell’s specific Claude-to-WordPress workflow for building rankable websites without relying on AI hosting platforms?
\n<\/h3>\nCraig’s workflow uses Claude as the code generation engine rather than as a hosting or deployment platform. He prompts Claude to write the site code, then manually copies and pastes that output into Gutenberg blocks on a WordPress installation. He hosts these WordPress sites on WPX Hosting, which gives him full server-side control: he can configure caching, speed optimization, and server location to match his target geography, none of which is possible when a site is locked to a Manis or Lovable subdomain. He also noted Elementor and Cadence as viable block-based alternatives to Gutenberg for this paste-in workflow. The MCP (Model Context Protocol) connection directly into WordPress was flagged as a potential accelerator for this process, but Victor’s observation that MCPs consume disproportionately high token volumes led Craig to continue with the manual copy-paste approach while his test sites accumulate ranking data before he commits to scaling the method.<\/p>\n
<\/p>\n
\n <\/p>\n\n
Authority Engineering<\/span> <\/div>\n<\/p>\n
\nBuild the Content Architecture That Gets Cited by AI Engines
\n<\/h2>\n<\/p>\n
AuthorityRank engineers expert-level content at scale: structured for LLM retrieval, optimized for AI-powered SEO, and built to establish your brand as the definitive source in your niche. Thirty publication-ready articles in five minutes. Citation-worthy by design.<\/p>\n