Ai

“The next set of winners will be firms that capitalize on service-as-a-software, i.e., taking human-intensive services and putting a thick layer of AI on top to scale with less labor. This is a fancy way of saying there will be more consumer-facing AI applications.”

“The AI Company of 2025: Meta. No business is better positioned to register progress in AI than Meta.”

“The next set of winners will be firms that capitalize on service-as-a-software, i.e., taking human-intensive services and putting a thick layer of AI on top to scale with less labor. This is a fancy way of saying there will be more consumer-facing AI applications.”

“The AI Company of 2025: Meta. No business is better positioned to register progress in AI than Meta.”

“Assuming there will be a 4.5x increase in worker productivity through 2030, and that the vendors of AI software can capture 10% of the value their software generates, ARK estimates the total addressable market of AI applications will be $13trn in 2030.”

“the costs of training AI models are expected to fall by as much as 75% annually through 2030.”

“So, like you said, we've also shared this view that next year's going to be positive, albeit there's going to be more volatility. And when I think about the two main risks that retail investors are facing today, one of them is definitely inflation. We're seeing that in services. We're seeing that in housing. We've had the labor market shrink over the recent couple of quarters, so who knows if wage inflation pops up again. But there are ways to definitely hedge against that in an equity portfolio. We think, for instance, owning parts of the AI infrastructure cohort is one of the ways of hedging, whether that be in utilities, pipelines, energy infrastructure in general. These are areas that we think are a necessary hedge against inflation risk. And number two are a positive diversifier.”

“AI training performance will increase 5x in 2024, thanks to the effects of Moore’s Law ..., as well as the fact that Nvidia has previously outperformed the gains predicted by Moore’s Law, and algorithmic optimisations such as Chinchilla Optimal Scaling.”

“I put forward a few years ago that actually OpenAI could stop making models and be one of the most valuable companies in the world just because of ChatGPT. That’s their most valuable asset. And part of the problem that they have is that was definitely the case in 2023 and 2024, but they never… You have to actually build the business model around that.”

“The rigid agency model is crumbling. By 2026, success will not be measured by roster size or retainer value, but by speed and adaptability. Expect the rise of fluid all-star team structures, curated specialists assembled per project rather than permanent departments. AI will accelerate this shift, enabling leaner operations. The winners will be agencies that treat flexibility as a strategy, not a compromise. Those clinging to bloated teams and inflexible contracts will struggle as clients prioritize partners who can pivot with market changes, scale talent as needed and deliver specialized expertise exactly when required.”

“The agency business model is imploding. Agencies are trying to abandon traditional billable-hours structures and pivoting hard toward productized solutions, selling packaged platforms, technology partnerships and performance guarantees instead of just creative services. It's the only way to charge for value created in an era of AI-driven process automation efficiencies, and because CMOs are demanding transparent, measurable ROI on every marketing dollar. Vague metrics and fluffy brand promises won't cut it anymore. In 2026, if you can't prove tangible impact with hard numbers, you won't get the budget.”

“AI is expected to be nearly fully intertwined within the digital advertising economy by 2029, when GroupM expects that 94% of ad revenue will 'be informed' by the technology. That figure includes programmatic, search advertising, and use cases like preplanning and ad optimization... 'After doing this research, there may be no other industry right now where AI is used more pervasively.'”

“Robotaxi platforms — the convergence point of EVs, AI and robotics — could generate $28trn in enterprise value over the next five to 10 years.”

“A lot of the output is trending toward the median. It’s about pulling against the median because all of the content is merging to look very, very similar.”

“And at least one company that’s tried and failed to get GenAI to work will vocally give up on it in a front page news story.”

“inventory scoring plus creating packages that are truly changing in real time and adapting is where the ad curation AI future lies.”

“it's really surprising, but people always know when it's AI.”

“My thought on on AI replacing people is that people that use AI well will replace people that don't use AI well.”

“OpenAI will launch ads in 2026.”

“In the coming years, we'll see a shift in focus towards applications that can handle complicated tasks based on our stated intent — eg, planning a trip to Italy.”

“Ads on ChatGPT will become CPA.”

“I'm pretty sure that the Agentic version of the market that is coming will be impression level, will be super fast, and will be born in this protocol.”

“I think one of the predictions I'm making for next year is that the AI D2C bubble is gonna burst a bit. But that AI, a prediction I made for this year, was actually going to be a very unsexy success. And one of the great examples of this is AI operating behind the scenes, not right overtly out there, but just creating greater discoverability, much more personalization, and in particular, going to solve some of the biggest problems in the ecosystem.”

“Confession: I predicted last year that Judge Brinkema in Eastern District of Virginia would issue her remedies opinion by end of February in U.S. v Google. To force divestiture. That’s in the next 24 hours. Polymarket has my odds super low. Alas.”

“by the end of 2026, brands will be able to simply tell Meta their objective and budget, and the platform will handle everything else.”

“Actual risks from modern AI. $1.5T in existing investment grade bond debt for AI data centers. Then there is training debt. And no cash flow positive products deployed. The more all in on new AI that a company is, the more risk. Failures will ripple through the economy. www.fool.com/investing/20...”

“Why Creativity Will Matter More in An AI Age New Canvas Human Voice Human Choice Plumbing and Poetry Conversational Interfaces rishad.substack.com/p/why-creati...”

“Adulthood is just cycling between "If I don't say yes to every opportunity, no new opportunities will ever be offered to me" and "Oh no, I have said yes to too many things" over and over until you die.”

“A delayed response to a @caseynewton.bsky.social segment from Hard Fork last week: Casey’s right, I think, that the AI bubble is more like railroads than crypto. But we should also be thinking about broadband in ‘00 and housing in ‘08. davekarpf.beehiiv.com/p/what-sort-...”

“I put forward a few years ago that actually OpenAI could stop making models and be one of the most valuable companies in the world just because of ChatGPT. That's their most ... valuable asset. And part of the problem ... that they have is that was definitely the case in 2023 and 2024, but they never… You have to actually build the business model around that.”

“More than one-third [of CMOs] expect to cut staff over the next one to two years”

“In 2024, artificial intelligence and the oceans will emerge as new issues motivating geopolitical competition and regulatory dynamics.”

“I would say, more likely than not, 50% of all Super Bowl spots we watch this year will utilize generative AI in some facet.”

“By 2028, brands’ organic search traffic will decrease by 50%+ as consumers embrace GenAI-powered search (70% trust from survey ~July 2023).”

“By 2025, 50% of consumers will significantly limit or abandon social media due to perceived decay in quality, including GenAI integration (70%+ see harm to user experience).”

“A subsection of brands will shun AI and prioritise more human positioning. This ‘acoustic’ concept will be leveraged to distance brands from perceptions of AI-powered business as impersonal and homogeneous.”

“The use of GenAI in a creative team’s routine daily work frees them up to do higher level, more impactful creative ideation, testing, and analysis. As a result, creative will play a more important and measurable role in driving business results, and CMOs will actually increase their spending on creative and content.”

“Gartner predicts that by 2026, traditional search engine volume will drop by 25% as market share shifts to AI chatbots and AI-driven search. Additionally, Gartner forecasts that by 2028, organic search traffic will decrease by 50% or more.”

“AI will change the way that shoppers find products. Tools like NLP-powered chatbots, visual search, and better recommendations will mean that in the future shoppers will be directed toward a smaller set of highly relevant products.”

“Increasing prevalence of AI-based activation” in retail media, led by Amazon and mature networks; retailers will automate audience/placement selection for outputs like ROAS or new-to-brand growth, balancing automation with advertiser controls.”

“AI is going to be increasingly important – both generative and predictive. [Epsilon] has been using predictive AI to get the right message in front of the right consumer at the right time for over 10 years. This is not going to stop and will become increasingly important to help drive the best outcomes for consumers, brands and retailers.” (On generative AI: “This too will be important to help manage campaigns and manage creative within retail media.””

“The agency holding companies are running out of organic growth options — expect a wave of mega-mergers as they try to match Google and Meta's scale on data and AI.”

“AI chatbots will disrupt Google search advertising. Not in 5 years — sooner. The behavioral shift is already happening.”

“AI-generated misinformation will define the 2024 election cycle in a way that makes 2016 look quaint.”

“By 1998, a computer will defeat the world chess champion.”

“On the highway towards Human-Level AI, Large Language Model is an off-ramp.”

“Auto-Regressive LLMs are doomed. They cannot be made factual and non-toxic.”

“People should stop training radiologists now. It's just completely obvious that within five years deep learning is going to do better than radiologists.”

“I set the date for the Singularity — representing a profound and disruptive transformation in human capability — as 2045.”

“The cloud and mobile are not the next big thing — they are the foundation. The most important companies of the next decade will be built on top of them.”

“By 2029, computers will have human-level intelligence and pass the Turing test.”

“Mustafa Suleyman Leaves Microsoft... Suleyman is an impressive, driven leader, but it’s fair to wonder whether Microsoft actually just needs an integrator to build OpenAI’s latest technology into its products. Already, reports have indicated that Suleyman’s team has clashed with OpenAI, and it wouldn’t be a shock if he exits in 2025.”

“A Prominent Company (or Two) Will Loudly Give up on Generative AI... at least one company that’s tried and failed to get GenAI to work will vocally give up on it in a front page news story.”

“In the US, activity should stay robust, underpinned by strong AI-related capital expenditure, still-easy financial conditions, and a positive fiscal impulse from the 'One Big Beautiful Bill'.”

“We expect global growth to remain robust in 2026, underpinned by reduced trade-policy uncertainty, a supportive fiscal stance, accommodative financial conditions, and sustained AI-driven capital investment.”

“We expect AI spending will exceed consensus estimates this year but begin to decelerate in growth terms while corporate adoption increases, causing rotations among the largest US tech stocks that create two-way risk for the aggregate index.”

“Hyperscalers capex is expected to reach $540 billion in 2026, with Nvidia’s CEO Jensen Huang anticipating $3-4 trillion of AI infrastructure spending by 2030.”

“Healthy economic and revenue growth, continued profit strength among the largest US stocks, and an emerging productivity boost from AI adoption should lift S&P 500 EPS by 12% in 2026 and 10% in 2027, providing the fundamental base for a continued bull market.”

“My prediction is that with both theories gaining followers and acknowledgment, they may actually merge in 2025 into something new. In my mind, they are complimentary to one another. If you think about it, if we are living in a simulation run by some kind of computer, then it stands to reason that there could be millions of other simulations running in parallel to ours.”

“Human-operated organic networks are being displaced by AI-operated inorganic ones, which 'are always on, and therefore they might force us to be always on, always being watched, always being monitored.'”