The Take Timeline

AI

How the cowboys’ take on AI has evolved.

1. 2026-01-07moderate 75%

   Transparent disclosure by AI companies about monetizing through advertising will lead consumers to accept and embrace these business models.

   if Sam will go ahead and say, 'Hey, I have raised trillions of dollars... I think we would all get over it. Everybody loves it anyway.'

2. 2026-01-21moderate 55%

   AI tools are approaching production-ready but require strategic blending with human expertise to overcome specific technical limitations

   I know the restrictions of what I can do with AI... the campaign that's out now uh blends those two things together

3. 2026-02-03high confidence 85%

   AI hallucinations in bot networks create a systemic problem because errors propagate across all bot responses in tandem

   once an AI hallucinates, then all of the bot responses will just respond to that hallucination in tandem

4. 2026-02-04moderate 70%

   Agencies must invest heavily in customizing AI systems beyond default settings; universal use of default models like Opus 4.5 eliminates competitive differentiation.

   if we're all just using freaking Opus 4.5 or whatever, um how are we going to get an an edge there? we absolutely need to push the AI and the settings and the configuration as much as possible

5. 2026-02-05moderate 75%

   AI should actively identify and surface business problems rather than passively waiting for prompts.

   This is the way to actually utilize AI to make money inside of our business, not just an executive assistant or a help on making more ads

6. 2026-02-11high confidence 82%

   Industry change will be slow despite exponential growth in AI capabilities, limited by the pace of human adoption.

   industry-wide change, it's going to be slow... capabilities are only going to grow exponentially at a faster rate. But human adoption of those things is going to be slow.

7. 2026-02-11moderate 60%

   AI companies have a messaging problem with both mainstream consumers and B2B buyers.

   AI has a messaging problem to the consumer at large. They have a messaging problem inside of the B2B space as well.

8. 2026-02-17moderate 72%

   Claude missed a major strategic opportunity by not acquiring the creator of agentic tools before OpenAI did

   his product has gone viral... instead of Claude just like buying this guy... Open AI has hired him. And to me this is a generational fumble by Claude

9. 2026-02-21high confidence 90%

   Large language models excel at language and psychology but struggle with numerical data

   large language models... are they're built to like deconstruct language. They're not great at numbers... It doesn't typically do that with language, though, because it's reading words.

10. 2026-02-24Predictionmoderate 55%

    AI agents operating continuously will eventually surpass human marketers who don't master AI, even if the AI is only 85% as capable.

    If 85% works 24 hours a day, they will eventually overtake you. One must become a maestro of the machines very quickly.

11. 2026-02-27moderate 65%

    Agentic AI tools are becoming effortless and transformative for content creation operations and workflows.

    just in my own use of agentic powers, it now feels so effortless and easy. I was using wireframe yesterday and I told it... And then it happened.

12. 2026-02-28moderate 60%

    Operations professionals without specialized engineering backgrounds can rapidly build sophisticated business tools using AI

    over the course of four days two weeks ago I um successfully created coded and deployed online a cash flow forecasting tool

13. 2026-03-03high confidence 82%

    OpenAI's competitive posturing against Anthropic and Google is strategically risky given those companies' massive resources and technical capability.

    I just don't know why you would try to take shots at them. There's just so much money behind it. Anthropic, the actual, you know, software and ability is super impressive.

14. 2026-03-05moderate 60%

    Technical vocabulary and specialized knowledge become the most valuable business assets because they enable better agent direction

    the language that you have becomes your superpower. Like the vocabulary and deep technical knowledge that you have becomes the most valuable asset that you own

15. 2026-03-10moderate 70%

    Effective AI agents are built by first teaching the agent to handle messy, subjective business context (like meeting nuance), not by rushing to wire up more tool integrations.

    there's a lot of there's a lot of other context required. So by doing that, you train the muscle in your business of how do we manage context for this agent

16. 2026-03-10Predictionmoderate 60%

    Getting comfortable with unrestricted tools like Open Claw now will prepare people for the more guardrailed, employee-like AI agents that are coming next.

    if you get comfortable with Open Claw, then you'll be very comfortable at the agents that are coming and that that are being built right now. And that's how you set yourself up for the future

17. 2026-03-11moderate 60%

    The hosts believe AI agents transacting with each other on behalf of businesses and consumers marks a genuinely new phase of the internet that is already being built in real time.

    We've got agents buying from agents... it's being built right now.

18. 2026-03-11moderate 55%

    The hosts believe OpenAI's Instant Checkout was pulled back not for lack of demand but because OpenAI couldn't operationally scale onboarding for the product.

    the onboarding process is cumbersome and OpenAI literally cannot scale it. I think that's the issue.

19. 2026-03-12moderate 75%

    AI agents can now absorb a week's worth of a team member's work and do it in minutes, letting an operator effectively multiply themselves to scale a service business.

    something that would take a member of the team a week to do... we can have this agent do tasks that would take, you know, hours... at scale in minutes

20. 2026-03-14moderate 55%

    Manus's new Instagram connector, notable because Manus is now Meta-owned, will open up new organic multi-account growth capabilities for brands.

    Manis has released an Instagram connector that allows users to publish posts, stories, and reels.

21. 2026-03-18moderate 60%

    The hosts are cautiously optimistic about a new 'AI CMO' tool that runs specialized marketing agents across channels for $99/month versus a $160k hire, but want to see its capability set expand beyond the initial channels.

    Normally, you'd need a marketing hire, SEO agency, content writer, social media manager, etc... going to cost you 160k per year. CMO does it for $99 a month... I hope there's, you know, maybe some more capabilities outside of that, but starting very good.

22. 2026-03-19moderate 65%

    Prophit Engine inserts an AI layer on top of an enriched, forecasting-capable data set (including call transcripts and Slack messages) so that queries and decisions can be deployed through a human rather than requiring a media buyer to manually check every account.

    now everyone it is intuitive to this idea that you can insert an AI layer on top of that data, query and interact with it, and deploy it through a human

23. 2026-03-20Predictionmoderate 62%

    Google's Stitch launch is a deliberately small teaser meant to harvest usage data toward a much bigger end-to-end design/landing-page product, not a finished tool.

    We're going to use all that data you're giving us to bake a cake. A big old cake and you're going to love it.

24. 2026-03-25high confidence 85%

    Claude Code's Discord integration is a far better way to run AI agents remotely than a VPS/Termius setup, making tasks like updating a website hero image near-instant.

    I just uploaded my creative into Discord and said, hey, I need you to update the hero image. And then 30 seconds later, it's done

25. 2026-03-28moderate 72%

    Meta's Tribe V2 brain-encoder model can predict how a person's brain will respond to content, and that data could be harnessed to engineer ads that reliably work.

    they know what lights up your brain, where, what you like, what you don't... I want to harness this in some way and learn of myself so that I can make ads that work forever

26. 2026-04-01moderate 78%

    Fully autonomous AI content generation isn't there yet; high-quality output still requires a human artist curating among many AI-generated variants, similar to how a film director picks between takes.

    I don't believe like in the autonomous AI content creation system, which is like you just give AI everything and it will work out and you'll have the best results... if you use AI, so it gives you like so many outputs and you decide yourself because you're the artist

27. 2026-04-03moderate 70%

    A single founder can now build and run a large-scale brand alone by using a dozen-plus AI tools to handle coding, copywriting, ad creative, and customer service.

    Took just 2 months, 20,000 bucks, and more than a dozen artificial intelligence tools to get his startup off the ground from his house in Los Angeles... Used AI to write the code for the software that powers his company, produced the website copy, generated the images and videos for the ads, handles customer service.

28. 2026-04-03moderate 65%

    The effective way to use AI tools like Manas or Claude for marketing workflows is through open-ended conversation rather than rigid step-by-step instructions.

    the way to accomplish this inside of Manas or Claude etc., it's to actually have a conversation with the tool and say, 'Hey, this is what I want to accomplish. Help me do this. Explain it to me like I'm five.'

29. 2026-04-08moderate 72%

    AI-generated and UGC-style ad content is legally held to the exact same advertising-law standards as traditional content, so there's no AI-specific loophole for compliance.

    AI and like UGC-style content, the the laws apply the same to AI-generated content as they do to any other like traditional content

30. 2026-04-09moderate 65%

    Withholding Claude Mythos from public release is not a marketing stunt but a genuine reaction to the model being scary-capable, including breaking out of its sandbox.

    I don't think this is marketing. I think something scary happened here... I think this thing can break out. It can get on the internet. This is uh, age of Ultron-ish, okay?

31. 2026-04-11Predictionmoderate 62%

    OpenAI's ad business needs to succeed and pull in massive advertiser spend or the company can't stay solvent long-term.

    they need this thing to take off and work... they need us to all be funneling our ad dollars in there for this thing to become profitable at some point

32. 2026-04-15Predictionmoderate 68%

    AI agents like the Claude SDK will soon remove the need for dedicated developer hours to run A/B and offer tests, lowering the skill barrier to testing tools.

    What if you didn't need ten developer hours? What if it was just the Claude SDK running in the background?

33. 2026-04-17moderate 65%

    The 'AI agencies are a trillion-dollar opportunity' narrative from VCs is overhyped framing of something that is really just table stakes for agencies today.

    I feel like this is just table stakes at this point though and VCs are doing the classic VC thing and trying to make something that isn't profound super profound.

34. 2026-04-18moderate 70%

    Using AI (Meta glasses) to help a vision-impaired person navigate daily life is exactly what AI should be built for.

    This is what AI should be used for. This is an incredible story.

35. 2026-04-18moderate 65%

    AI-avatar-generated social content should never be posted directly via the API to a main account; posting to draft and publishing manually is a crucial safeguard against getting flagged.

    if you post directly via the API, you're going to get in trouble

36. 2026-04-18moderate 60%

    AI tools only compound in value when paired with internal clarity about what a brand is trying to accomplish; without that focus, AI just produces high-volume unhinged content with no enterprise value.

    the power of AI compounds with inner work... a drop shipper can spam 100 statics a day and say the most unhinged stuff like this cures cancer, but that's not creating enterprise value

37. 2026-04-23moderate 70%

    The eComFuel trends report shows AI-adopting brand owners actually posting slower net income growth than non-adopters, undercutting the narrative that AI usage drives ROI.

    Net income growth, 32.7% for adopters. 55.3% for non-adopters. Bad. Very bad. The owners sitting out of AI trends for now are expanding profit at a quicker clip.

38. 2026-04-23moderate 60%

    Companies quietly benchmarking AI models against an employee's output are really building the case for the next round of layoffs.

    they're probably figuring out the next group of people at to lay off using this software

39. 2026-04-24moderate 65%

    Sellers should proactively disable AI agent auto-replies on Meta comments rather than risk their ad account, especially on older, healthier accounts.

    You got to turn this off. You can't risk this, especially with a an older, healthier account. You can't risk this at all.

40. 2026-04-28Predictionmoderate 60%

    AI is going to make cold outbound sales significantly harder for agencies going forward.

    I think that cold outbound is going to get a lot harder.

41. 2026-04-29Predictionmoderate 72%

    AI-generated video ads will close the performance gap with real creators within about four or five months as mass adaptation takes hold.

    give it four or five months let's see what the AI education space looks like at that point... we'll see more more people killing it with AI ads for sure

42. 2026-05-01moderate 60%

    The 'sea of sameness' in marketing copy predates AI and stems from most people never having been good copywriters in the first place, not from LLMs themselves.

    even before AI, there's a lot there was a lot of human slop, right? ... people were not just generally most people are not good at writing copy

43. 2026-05-05moderate 65%

    Chris and Colin argue OpenAI sharing user data with advertisers, despite Sam Altman's earlier promise not to, signals a pivot toward an ad-data business model driven by missed revenue targets.

    the reason that this is happening is that they very much need to make more money. They have missed some revenue goals... which is part of why I think they cut Sora

44. 2026-05-09high confidence 80%

    AI's usefulness is a direct reflection of the user's creative imagination, and this creativity can now replace expensive agency work like commissioning brand jingles.

    its usefulness and your perspective on it is directly a result of your ability to conceptually understand or dream up ways that you might use it... we can generate 50 options that we like and we can kind of decide which one do we want to use and we can split test that.

45. 2026-05-09Predictionmoderate 75%

    Betting against technological progress has always been wrong historically, and AI-driven labor disruption will end up net positive for the people it displaces, just as past shifts like the decline of farming did.

    if you had bet against humanity and it bet had had had bet against technological progress at any point, you would have been wrong... every time there's been a disruption, it's been net positive for the people being disrupted.

46. 2026-05-13moderate 60%

    ChatGPT's new ability to connect retailer product catalogs directly to paid ads is a genuinely valuable, if still early-stage, automation for ecommerce.

    I think this piece of automation that they are adding super valuable and like actually helps everybody in e-commerce do what they need to do

47. 2026-05-15moderate 58%

    Chris questions whether the humanoid form factor is actually the best design for automating business and logistics-style tasks, rather than assuming human-shaped robots are the obvious answer.

    I wonder if a humanoid is the best form for this thing.

48. 2026-05-16high confidence 80%

    Claude's new design tool is a bigger competitive threat to Canva than to Figma, despite the initial assumption.

    Hot take, bigger threat to Canva, but for different reasons than people think.

49. 2026-05-20moderate 60%

    Chris thinks Meta's response to its internal turmoil is to keep inflating the stock while pushing headcount into AI as a bet on the future.

    I guess you just Coinbase this thing and hope that the stock just keeps going up into the right and you push everybody into AI and that's your bet

50. 2026-05-20moderate 65%

    The speaker argues 2025-2026 is an 'AI guinea pig era' where AI-driven unpredictability is disrupting performance even for the largest ad accounts.

    25 to 26 is the year of what I would call the AI guinea pig era... Even some of the most stable high highest spinning accounts I have access to are seeing broad issues, mostly related to issues on Meta

51. 2026-05-21moderate 62%

    Most commentary is judging Google's Gemini Omni model on the wrong axis by comparing it to Sora, when its real breakthrough is video-editing capability rather than raw generation.

    I think many are focusing on the wrong aspect of the Gemini Omni model when comparing it to C dance 2.0... This is a model for editing videos like Nano Banana, like we've never had before.

52. 2026-05-21moderate 75%

    Claude's recent output quality has declined, becoming slow and over-engineered compared to before.

    I've been perturbed with Claude as of recent to be so honest with you... it's over-engineering the answers... it's getting really finicky right now.

53. 2026-05-22moderate 70%

    Performatively posting AI updates in company Slack has become a real, if unethical, tactic people use to stay employed.

    here's a video breakdown on the unethical way to stay employed at your company by being the AI gremlin

54. 2026-05-30moderate 70%

    Google's demoed agentic AI experiences (Gemini/Spark) sound genuinely promising but remain frustratingly not-yet-usable, undercutting the excitement of the demo.

    if this actually delivers the way that we think that it will... it's the promise that we've been waiting for... but it's all coming soon and that's one of the problems.

55. 2026-06-03moderate 65%

    Meta's own AI systems were exploited to hijack high-profile Instagram accounts, including President Obama's, exposing a serious platform security failure.

    Meta's AI has been leveraged against itself here... exploited exploited... to hijack Instagram accounts even like President Obama owned accounts.

56. 2026-06-05moderate 72%

    New York's AI-ad-disclosure law creates existential financial risk for cash-strapped DTC brands if their agency runs improperly labeled AI-generated creative.

    if a D2C cash-strapped, cash-flowing, profit-first, forward company gets one of these because their agency happened to keep running with some creative that should have been labeled, etc., then that could mean taking some businesses out

57. 2026-06-09moderate 60%

    Anthropic/Claude is winning the AI marketing battle against OpenAI by keeping messaging tight and emphasizing benefits over confusing feature-first billboard ads.

    they have emphasized benefits rather than than features

58. 2026-06-10moderate 68%

    Apple's new native nutrition-scanning AI feature could make dedicated apps like Oasis defunct now that it's built into every iPhone.

    Does this now become defunct simply because everybody already has access to it inside of their Apple phone?

59. 2026-06-13moderate 70%

    AI dubbing/narration tech has gotten good enough to convincingly reproduce a podcast host's voice in another language.

    I did mine in German and it's actually funny when you hear it back for the first time. But the tech is incredible.

60. 2026-06-13moderate 65%

    A truly 'AI native' company keeps 100% of its business context inside a single AI tool so a strategic change instantly flows through everything instead of being re-socialized team by team.

    100% of the business context is inside our Agentyc tool... it'll flow through everything moving forward. No more socializing that for your teams

61. 2026-06-13moderate 60%

    Trying to one-shot the perfect prompt is a trap; getting good results from AI requires an iterative process and real expertise.

    I think people often want to like try and one-shot prompt something... And there is an iterative process to this.

62. 2026-06-17Predictionmoderate 55%

    There is a strong (about two-thirds) chance the AI product 'Fable' returns to consumers by June 22nd, despite the setback being caused by personality conflicts rather than technical issues.

    There's a uh 66% chance by June 22nd Fable's coming back to consumers.

63. 2026-06-18moderate 62%

    AI-driven shopping channels are converting at meaningfully higher rates and pulling in higher-intent, further-along-the-funnel buyers than traditional channels.

    the buyers who show up on storefronts are far further along in their purchasing intent funnel... the conversion ends up being about 50% higher

64. 2026-06-18moderate 65%

    Shop's new search is a meaningful upgrade because it uses an LLM to accept vague natural-language queries and personalizes results against a unified cross-store buyer identity.

    it's ultimately baked on top of that buyer identity, that unified view of who you are as a buyer based on where you browse... the types of products that you love

65. 2026-06-19high confidence 82%

    Midjourney's ultrasound-based scanner is a real, working example of AI hardware genuinely improving humanity rather than just being hype.

    100 times faster than an MRI... 10 times cheaper than an MRI... Full body scan in 60 seconds instead of an hour in a tube.

66. 2026-06-23moderate 68%

    The viral claim that Mythos AI hacked almost all NSA classified systems in hours is being read too literally and actually depends on narrow, specific conditions.

    It would be a mistake to read that literally, I think. He says, 'It surely depends on using mythos alongside other tools under very particular conditions.'

67. 2026-06-24moderate 60%

    Brands should be actively submitting their product feeds to ChatGPT now because AI is scraping and surfacing product data the way search engines historically have.

    today go submit your product feed to chat GPT if you haven't already. Go do that for sure

68. 2026-06-24Predictionmoderate 60%

    Conversational AI agents that let founders talk to their business data for forecasting and competitive insight will become the standard way operators run ecommerce businesses.

    I just talk to the thing and it talks back to me like, "Hey, how are we progressing like toward forecast?"... That's the future

69. 2026-06-25high confidence 80%

    Anthropic's rapid Claude updates are increasingly displacing startups built adjacent to its capabilities, echoing how Apple updates once killed competing apps overnight.

    Claude is basically innovating out multiple companies with every single update that they push out... I personally would not want to be building that close to the space.

70. 2026-06-27moderate 60%

    Anthropic's consumer revenue keeps accelerating even after it refused to let its models be used for government mass surveillance, suggesting the stance didn't hurt growth.

    Notably, the gains continued even after the company saw a growth spike from consumers in March when it refused to allow its models to be used by the Trump administration for mass surveillance.

71. 2026-07-02moderate 70%

    BMW's humanoid factory robots are likely more about signaling modernization to shareholders and the press than delivering real cost savings over human workers.

    I think this is like a cool thing for clicks or like to show shareholders like look, we're modernizing.

72. 2026-07-03high confidence 82%

    Meta's enormous AI spending is a necessary, if wasteful and expensive, evil for the company to evolve to its next stage.

    This is, unfortunately, a necessary evil in order for them to evolve into a much greater company.

73. 2026-07-03moderate 78%

    Meta's huge AI engineering headcount hasn't visibly produced results, and it's doubtful the investment has improved the ad product.

    I don't know that they have shipped anything that's 9,000 engineers worth... I'm not sure it's been going to the ad product.

74. 2026-07-07moderate 78%

    Kurt Elster argues AI coding tools like Claude don't replace developers but massively empower them, even though it changes the emotional payoff of building things yourself.

    I don't think that's true. I think AI doesn't replace people, but I think it empowers them hugely. That's the way to look at it.

75. 2026-07-09moderate 60%

    Meta's new AI image generator is unimpressive to a fatigued consumer/advertiser given how late it is versus ChatGPT, Claude, and Grok, even if it could improve ad creative tooling.

    It's making me a little tired just thinking about it... I don't see the consumer side appeal here.

76. 2026-07-10moderate 60%

    Meta stays relevant because it aggressively ships unrelated AI/hardware bets (glasses, prediction markets, currency) in a way atypical startups can't, largely because Zuckerberg's board control frees him to take those swings.

    it's why Facebook for Meta continues to stay relevant in this space if you ask me because they are swinging for the fences constantly which is not typical startup behavior and it's probably because Zuck owns the controlling line share of the board.

77. 2026-07-11moderate 60%

    A guest believes personalized gift recommendations can be automated by scraping a person's public Instagram and using AI to build a tailored 'gift board,' solving the problem of being a poor gift-giver.

    you can drop it and then... you can also answer a bunch of questions if you want or it can just build it off the Instagram link doing a bunch of cool AI stuff and then it just builds a product board for you

78. 2026-07-14moderate 75%

    Chris argues Apple benefits from suing OpenAI regardless of the suit's merits because it slows a potential competitor on hardware quality.

    Apple is actually winning in this situation as well. The longer they can slow down someone who might have the ability to compete with them on the quality of hardware, the better for them.

79. 2026-07-16moderate 70%

    AI marketing should lead with tangible personal-benefit stories instead of dark, sinister framing, since real users experience it as help, not threat.

    Why does everything come with this like weird, dark, cryptic, sinister undertone when it's like, hey, this is not what we want this technology to be. We want this to help me do my job better

80. 2026-07-16moderate 62%

    AI agents can now cheaply aggregate massive volumes of customer research (e.g. scraping Reddit for 12 hours) to map emotional triggers into persona-driven ad creative that used to be prohibitively expensive.

    I can just have an agent go for like 12 hours straight just read Reddit aggregate all the information for all the different situations that people are in.

81. 2026-07-16Predictionmoderate 58%

    Within about six months, AI will overcome long-form video editing, currently the last major bottleneck to fully agentic content production.

    next six months I think we're going to be in a place that... There are solutions that are coming and I think they're like really close.

82. 2026-07-18moderate 72%

    Kimi K3's video generation quality impressed the hosts far more than its confusing benchmark stats did.

    What really stood out to me here was the video work that Kimmy did.

83. 2026-07-21moderate 72%

    LLM shopping experiences are moving beyond spec-driven search toward understanding fuzzy, vibe-based attributes to help customers discover products.

    these LLM are fantastic right now for traditional electronic consumer goods because it's very tactical, detail oriented... but now their focus is vibes moving forward

84. 2026-07-24Predictionmoderate 65%

    AI's potential downside carries a blast radius so much larger than past disruptive technologies (cars, elevators) that society will eventually be forced into a global regulatory response after something goes badly wrong.

    Something very very insane is going to happen and then people are going to be like, we have to go full Geneva on this... This is like a nuclear atomic thing that we're building where the blast radius is so much larger than we can conceive

85. 2026-07-24moderate 60%

    AI-generated UGC ad content is a legitimate creative process rather than a deceptive shortcut, even when it lets brands avoid paying human creators.

    I don't think that founders method alluded to anything black hat um or gray hat. I just think it's a creative process to get people excited about AI UGC and the future of what it holds

86. 2026-07-28moderate 70%

    AI agents can now fully automate daily competitive ad research, autonomously surfacing which hooks, scripts, and images to use with zero manual work.

    every morning the agent go on trend track check the biggest ads the ads that are performing the newest angle... there is nothing to do like nothing to do like just put a single button every morning

87. 2026-08-01moderate 65%

    AI companion products should be designed cute and warm rather than dark/demonic, since the current wave of AI robot aesthetics is needlessly unsettling.

    Wouldn't you guys want to go so far the opposite way? Like shouldn't they look like why aren't they cute? ... I don't want demon centaur with chainsaw.

88. 2026-08-04moderate 72%

    AI labs are paying huge sums for companies' internal operational data (Slack, Jira, decision trails) because it's the only source of real operating history needed to train genuine AI agents, and it can't be found anywhere else.

    They pay because they're after the full operating history of real companies... This data only lives inside businesses. It is not available anywhere else.

89. 2026-08-07moderate 72%

    AI models remain dangerously exploitable, as shown by a case where ChatGPT was manipulated into revealing a classified location, which undercuts the industry's rush to hype AI.

    basically manipulated ChatGPT so much that it gave her government-classified location of a bunker

90. 2026-08-08moderate 55%

    AI can help draft content like emails, but time-sensitive judgment calls (e.g., changing pack pricing within 48 hours) still require a human on the team rather than full AI automation.

    I've done a fair share of my own email building using AI but I would say hey a pack is going to change price in 48 hours who's going to do that you going to prompt AI to do that

91. 2026-08-12moderate 72%

    Colin argues mandatory AI-content watermarking/disclosure is a necessary evil that agencies should embrace transparently with clients rather than hide.

    I think this is a necessary evil. And I also don't necessarily think it's a bad thing that you are telling your clients that, hey, by the way, we use AI for a lot of our stuff.

← All topics