I’ve already read too many “prediction” 2025 posts and we’re about to go into uncharted territory globally, so writing one doesn’t even make sense. I also had a chuckle at Ben Werdmuller’s personal 2025 OKR’s post intro about “This may be the most LinkedIn idea ever, but…”
I’ve never met Ben outside of online circles, but it was a more original way to look at the year differently. (Even if it was the most LinkedIn idea ever.)
My spin on this was to de-age myself by twenty years and do what some of my younger friends tell me is a sure-fire method: Manifest some shit.
Let’s manifest the shift we need to see in 2025.
We’ll finaly admit that the Algorithm Era failed.
Outside of maybe TikTok and Google’s iOS app, after fourteen years we’ll declare that algorithms have been a clear and spectacular failure. The measure of success for algorithms was to surface stories, video, and images that gives us a deeper connection to others, entertain us, or unlock a better understanding of the world at large.
That never happened.
Meta’s algorithms hurt Facebook’s use with Gen Z (dialing up negative sentiment and its devastating consequences to drive engagement versus surfacing your friend’s photos), Instagram (dialing up templated influencer content and devaluing friends), and Threads (week old content that was performing so poorly that they had to pivot in the face of a competitor). They were hardly alone.
Linkedin dialed up the signal on comment volume so much that AI bots broke the platform. Snapchat ramped up sponsored content to the point where Gen Z is using it mostly as a DM tool. And Twitter became an unusable, gated slop factory purely to boost Elon’s posts.
Reality needs to set in for them and for us. If your platform isn’t intended to be a discovery engine, it won’t ever be a discovery engine. Reels is learning that lesson the hard way, so much so that Meta’s PR team worked overtime over two years to plant stories about TikTok in the media. Rather than build on the principles of community and social media, they invested in marketing and PR exercises.
As we enter the quarter century year of this millennium, I hope we accept the reality that’s been in front of us since the Facebook feed went mainstream: social media can have unlimited potential but it needs active curation, customization, and moderation. By us — people and businesses.
Not platforms. It was a worthy experiment, but we now have a definitive answer and need to move on.
We’ll gatekeep and isolate AI from grifters.
After beta testing and building custom apps for AI platforms for several companies the last few years, I’m a defender of the ethical use of this tech. I’m also a vocal champion of the developers, data scientists, and companies leveraging the tech in the way Blockchain should have.
But I’m seeing a troubling trend that needs serious intervention: The same grifters who got into Bitcoin investing (rather than building something useful on the Blockchain) are turning their gaze to LLMs. In 2017, I took the attitude of “OK, let’s just watch them trip over themselves.” They did that.
What I completely missed was that they’d destroy confidence in the underlying technology, simply by associating themselves with it. The brunch lords sucked capital away from truly good uses of Blockchain (particularly as a ledger for authenticity or verification) and the conversation turned to Shiba Inu shitcoins or Hawk Tuah.
We can’t – I can’t – let that happen this time. I’ve sat in virtual and IRL meetings this year where well-meaning people are straight up guessing the future of AI technology without understanding the basics of its use today. (“AI is going to replace humans in their jobs next year in a big way.” My dudes — up until a few months ago it couldn’t count the R’s in strawberry.)
LLMs are experimental and not good enough at the moment (without fully being trained on clean and accurate data) to be a finished product. And that’s OK. I am and you should be fully aware of this when you pay a subscription. In my head, I know the value of what I’m paying for is actually computing power and not a fully baked product that will accurately deliver what I need. And that’s OK for me right now, an individual.
But businesses selling AI models to other businesses? That’s where the wheels come off. As of today, December 13th, 2024, the data across all AI models are unreliable, the privacy of your own data is suspect, the training sets are almost all illegal and open the door to litigation if you run images or videos in the market, the end user experience (especially for Chatbots and answering services) is on par or worse than automated phone service.
As I said, I’m a huge proponent of AI tech. I’m deeper into the use cases than almost everyone in my field (advertising).
But I’m also a major skeptic of the AI business in the near-term because of the insane valuations and current costs exceed its potential. I’m going to spend a lot of 2025 showing clients, colleagues, friends, and family the best uses of AI which are rooted in utility (coding, organizing information, building agents, finding lost digital things, combining unstructured data-sets) and not content generation (writing professional letters, responding to clients with an AI persona, commercialized images/video).
The potential for a profitable AI leader is certainly there. But it isn’t a reality now.
We’ll start to appreciate the things that work and we ignored.
I don’t have anything original to say about politics that hasn’t already been said. But on a higher level we’re about to see a lot of people take a painful crash course in the systems and departments that hummed away in the background and kept the trains moving (sometimes literally).
One of my all-time favorite lines is from G.K. Chesterton, a big voice in the Catholic literary revival in England. “Do not remove a fence until you know why it was put up in the first place.” We’re about to tear down a lot of fences with the overconfidence that only a failing world power can have.
For me, I have no control over that. What I will be doing is increasing my own investment and appreciation in what seems normal or “a given” on an everyday basis. Marriage, children, friends, family, jobs, books, health.
My hope for 2025 is you do, too.
We’ll stop confusing what news is and asking “Where do you get your news?”
There are only a handful of consistently strong news organizations. ProPublica. The Guardian. The Verge. The Texas Tribune. The Associated Press. Reuters. The Columbia Journalism Review. That’s the list.
I have a few journalists that follow me across platforms so before I take shit: Do you not see your publication on the list? That’s because the amount of articles you publish relative to the amount of investigative journalism you produce isn’t acting as a newspaper.
In 2025, we need dedicated editorial newsletters that have their own revenue stream. We deserve to see the end of paid newspaper columnists and celebrity editorials and have them replaced with credible voices we don’t hear. We deserve to see world-renowned experts in someone’s field replace Aaron Sorkin’s fever dreams in the New York Times. (And I fucking loooove Aaron Sorkin.)
I’ll pay you for this. Until then, you will get every Adblocker I can throw at you. (All of the above publications I’ve white-listed or pay for.) New York Times, Washington Post, LA Times, Boston Globe — they got soundly beat and lapped, and putting four Ivy league talking heads on stage for a fireside chat on “The Future of Media: Where We Went Wrong” won’t help.
And finally…
We’ll invest the time to read and meet the people who are getting “it” right.
We’ll stop reading Linkedin hot takes on social media. We’ll start reading Rachel Karten and Matt Navarra.
We’ll stop reading tech clickbait. We’ll start reading Nilay Patel, Dave Winer, and Ben Werdmuller.
We’ll stop replying to hustle porn influencers. Instead, we’ll dust off our Anil Dash, Priya Parker, and Seth Godin bookmarks.
I’m going into 2025 knowing what to do. And maybe, just maybe — if we do invest the time in people getting it right, if we do learn to separate news from views, if we do start to invest in the things that make us better, and if we isolate the grifters — the algorithms can’t do any more damage than they already have.
—
I’ll be making major updates to the site in the next few weeks. In the meantime, find me on Bluesky, Threads, and Mastodon.