/reader

Happy Ninth of Bleptember, Internet!

📰 Using a feed reader is the best way to read my blog posts. How clever you are to know that! 🚀

Gabe Rivera started this experiment working as a developer at Intel in the early aughts. He launched it as a full-time project 20 years ago this week – September 12, 2005.  Initially he called it tech.meme orandum. By March 1, 2006 it officially became what everyone was already calling it, Techmeme. I’m always looking for new and better ways to consume news. I was hooked the moment I saw it sometime late in 2005 or early 2006. I still am.  

“It chews through thousands of blogs in minutes and tells you what’s important. It does this every few minutes. It is dramatically faster than I could ever be. It’s all machine based. No humans involved,” said Robert Scoble when he broke the story back in 2005. He worked at Microsoft then and had one of the most influential blogs on the internet at the time, 

Ebay acquiring Skype was the top story the day Techmeme launched. It also listed commentary from half a dozen tech blogs including this “SkypeBay done for $4.1 Billion” from my partner Om Malik. Yes, both of us have been writing about tech for a long time. 

I’m telling you all this not just because Techmeme’s 20th birthday is on Friday, but because of how anomalous Techmeme has become among news websites. Very little about the way tech news – and all news, for that matter – is reported, edited, distributed, paid for and consumed bears any resemblance to the way it was in the fall of 2005. 

I now read almost everything on my smartphone. And instead of only reading specific publications that arrive on my doorstep or my mailbox, I subscribe to the digital versions. I read a few via their apps. But for most I allow a half-dozen social media channels along with Google News and Apple News to surface interesting stories.

I also subscribe to newsletters like someone with addiction issues. Om, my partner, thinks that we’re already living in a world where websites of any kind are irrelevant. 

But Techmeme looks and works exactly the same way as it always has. And it has never been more popular. Traffic is up 25 percent this year, likely driven by the explosion of interest in AI,  Rivera says.

Both he and I believe that what it does as an aggregator has never been more important given the explosion in sources of news and commentary online.

And while technically Rivera’s crawlers and algorithms don’t fit into the modern definitions of an AI system, journalism could learn a lot from what he’s created. Practically he’s been doing for 20 years what journalism outfits are only experimenting to do now with AI.  

It’s actually because of this consistency that it and its sister websites Mediagazer and Memeorandum remain one of the top three sources I look at every morning. (Rivera also publishes a celebrity news site called WeSmirch.) The only other website I can think of that has changed so little is Craigslist. 

Bill Gurley, the famous venture capitalist at Benchmark, perfectly described why he – and I – find it so valuable in a YouTube video with Tim Ferriss back in 2023. He said he uses Techmeme and Twitter together as a giant information funnel where they become a guide to all the important stories of the day “ Everything (how I read the news) has been inverted,  where (instead of starting with specific publications) the aggregators of Twitter and Techmeme are funneling what I find. Then I go read those things.”  

This is exactly why Rivera created Techmeme. ”I’ve always read a lot of news, and as a software person, I really found it cool that you could write crawlers to aggregate and mix the news in various ways. So when the blogosphere took off around 2002 it was really exciting for me because it created this real-time and opinionated layer on top of the rest of the news. And so what seemed like an obvious idea popped out to me at the time, which is just crawl a bunch of blogs, and then rank and surface what they’re linking to.”

One of the features that has always existed, but that I just learned about from talking to Rivera is that all 7303 days of Techmeme are easily accessible just by clicking on the date at the top. It brings up a box where you can enter the date you want to see and immediately displays that day’s top Techmeme headlines.  It’s a gold mine for Silicon Valley historians. I wish every news site had this feature.  

The site didn’t look like much in 2005. And it really looks dated today. There are no photos, videos, or graphics or sound. You get about 30 ocean-colored headline links and black subheds all against white background. Clickable arrows hide the most trafficked conversations on all the big social media networks about each headline.

Rivera, who is 52, said he picked the design entirely for readability and distinctiveness. It’s worked. Is there any other website out there that uses blue/gray/green as their primary color palette? “I probably wanted it bluish but also something distinctive and ended up there.”

It was designed for what he calls “information density,” not to be pretty. The goal is for it to load quickly on any device over any connection and be used quickly by busy people. 

“There’s a hierarchy of headlines. There are related headlines. They’re related social posts. We want it to be readable, scannable and we want it to also provide a visualization of the news where a bigger cluster is a bigger story. And we want to have both importance ranked and have there be a time sorted column of the newest stories on the right. So as long as we want to have those characteristics we’re locked into a certain type of design,” he said.

He said he gets asked all the time about when he’s going to add photos, graphics, podcasts and videos. He still gives the same answer: Techmeme is about what’s going on right now, and visitors need to be able to read it fast. That means every medium other than text slows things down. ”Either that kind of content is not timely or if it is timely, it’s not high density enough,” he said. 

Rivera hasn’t gotten rich doing Techmeme. But it hasn’t made him poor either.  “In a particularly strong advertising year I feel almost like a well paid media person,” he said. And it’s his. He hasn’t needed to borrow a dime or take a cent of venture capital to build it. 

Some of his operation’s longevity is because he still runs the technology on a shoestring budget. While most other websites and apps of substance have long since adopted cloud computing offerings from Amazon, Google or Microsoft, Rivera still runs Techmeme and the other three sites he publishes off two servers located in a single hosting facility. He said all the newsletters and a newer business for corporations generate between $1 million and $5 million a year.

Initially, it was just him running it out of his two bedroom apartment off Castro Street in downtown Mountain View. (He’s based in New York City now.) Back then the headlines were written and sorted entirely by software running on a server in his closet. He actually started the political version of Techmeme, Memeorandum, first while he was at Intel.  WeSmirch, the celebrity site, started in March 2006.

By 2007 he’d hired Omer Horvitz, a friend from grad school at the University of Maryland. By the end of 2008 he’d hired his first part time editor for Techmeme, Megan McCarthy, realizing that letting the technology do everything led to too many mistakes.  By March 2010 he’d launched Mediagazer, a headline site about the media business.

Now the software finds and ranks the stories. But the editors write the headlines. When stories are generated by corporate press releases/announcements, they choose which media outlet’s story is driving the most interesting social media conversations. The software also chews on the API feeds from the big social networks to come up with a list of the most useful conversations. Editors approve all those, however, to prevent duplication. 

He said Memeorandum and WeSmirch don’t get as much attention or traffic because they have no editors and because he hasn’t promoted them as heavily. He said he’s found it much harder to make money from them. 

Today the only full time employees besides him are Horvitz and two full time editors. The rest of the staff – a book keeper and the remaining 23 editors – are part time working remotely in enough different time zones to supply Techmeme and Mediagazer with near 24×7 coverage. 

Most of the revenue is advertising. But it now also includes fee-for-service revenue from companies, an offering that launched in 2020. Rivera describes it as news filtering and organization services for corporate customers. This often includes a custom Techmeme page customized for each company only with headlines relevant to them, their competitors or their industry.

It sounds like a more modern version of a clipping service, even though Rivera doesn’t think that’s the best term. Companies used to pay public relations firms or employees to constantly find and supply a list of stories being written about them and competitors.  “I shouldn’t comment on the number of customers. But it’s meaningful for us and the trends are positive,” he said.

He said because Techmeme is an aggregator rather than a content site of its own, it hasn’t needed referral traffic from Google. It’s therefore not been hammered by AI’s impact on publisher referral traffic. But he also knows that online advertising generally is in flux. It’s made him wonder sometimes if in the future he might offer subscriptions or develop other revenue streams. 

Like everyone he’s paying close attention to the AI boom. “We’ve only scratched the surface so far, mainly integrating LLMs to assist our editors with headline-writing. By carefully introducing intelligence at various points across the software system that handles our curation, I think our processes will get smarter and faster. And I would predict AI can especially improve the verticals with no editors on them like Memeorandum.”

Does Rivera worry that AI will eventually get good enough to simply supplant what he does, where all of us everyday will soon just be asking our favorite chatbot to list the top linked stories in our worlds? “ I can think of a lot of potential reasons they’ll come up short in the near term,” he said. That’s provably true. Ask ChatGPT to list the most linked to tech stories and all its sources are … Techmeme.

Rivera says that he’s not naive about the long term, however, “given the astonishing rate of improvement in AI capabilities we’ve seen. So we just have to improve our own stuff. And a major part of that will be adopting AI ourselves.”

Despite all this the basic approach of the original Techmeme algorithm remains the same, he said. “What are the most linked blog posts and news articles from this set of blogs? And once they reach a certain threshold, they’re featured on the site,” he said.

Maybe there’s a lesson here for the rest of the media world. I and every writer and mid level editor I know has stories about design changes to publications that made us groan. They seemed more in service of a new editor or design chief marking their territory like a dog or cat, than in service of actually making their publication easier to read.

Yes, it’s hard to do nothing. And yes, sometimes the best changes come from following the urge to tinker. But we can also learn this from Rivera’s two decades: More often than not, less is more. 

I’m still far from an expert, but my DIY abilities are growing with every project. My latest progression began with my bike: a simple disc brake repair, new grips, and a new luggage rack.

From there, it continued indoors. I faced a stubborn curtain rod on a difficult wall that was a mix of metal, concrete and porous plaster (we are living in a building from 1938). Today’s biggest challenge, however, was a completely rusted kitchen faucet that took three hours, lots of sweat, and a fair amount of stress.

But it got done. The real reward wasn’t just a new faucet; it was the confidence of knowing I can figure things out and get them done myself. And while I hope I don’t have to replace another faucet anytime soon, I know I’ll be ready when I do.

Interactions & Comments

It made me think of something I’d said here last month when discussing my pacific circuit.

Complicating the idea of my pacific circuit is that through either resilience or resignation, as my world shrinks due to the still-undiagnosed fatigue, I keep marking most days as having been “neutral” (my general target for overall state of mind) despite, for example, not having been to the zoo—a once weekly, then monthly, activity—since May and not having blogged—until right now—for two months, when those things were of a self-regulatory nature. It’s entirely possible that through such resilience or resignation I’d continue grading on this kind of curve when imprisoned in that room somewhere in New England, but as much as this dynamic makes me sad now it would only make me all that much sadder then.

What I’m stuck on in thinking about that bodymind tension is this question of resilience versus resignation when it comes to grading my days on this kind of curve. If I adjust my thinking and feeling to grade my days based upon the options of what’s currently being made available to me, doesn’t this suggest the disguising of an inevitable downward trend?

What if this sort of “grin and bear it” grading curve is a kind of masking, that thing autistics do that if maintained for too long and too deeply risks burnout. “Compensatory attempts are taxing,” I once quoted some researchers as saying, “need to be sustained over time, and are often unsuccessful, resulting in a cost to wellbeing.”

There was a Vice piece about resilience during the pandemic that I ended up separately citing here twice without realizing it. There’s a quote I used the first time worth bringing into the conversation here.

Resilience might sometimes look like grinning and bearing it. “But there’s times when resilience may look like crawling back into bed and crying,” Bedard-Gilligan said. “Feeling those emotions and processing through whatever it is that’s causing them. It may actually be the most adaptive thing you can do at that moment.” Bonnano coined the phrase “coping ugly” for the things we have to do in some situations to manage in the moment.

The challenge, however, comes in the calculus of which is more detrimental to your health: grinning and bearing it or “coping ugly”. You can’t crawl back into bed every single day.

Me, two years ago:

We have something of a botched view of resilience as being something we demonstrate through endurance rather than through recovery.

When I talked here about “grading on this kind of curve when imprisoned in that room somewhere in New England”, the reason why I see this as a fate akin to dying is that there’s be nothing in it but the enduring of it. There would be no hope of recovery, because the situation would never end, until, eventually and inevitably, I did.

Me, one and a half years ago:

The funny thing about to some extent reclaiming that mediocrity is that, as I’ve also discussed in therapy, whenever I’ve done any mood tracking, I’ve tended and trended toward weaving back and forth around a basic midline. In my original tracking app, I had this labelled “Meh”. Now that I’m tracking at the end of each day on my Apple Watch, I select “Neutral” more than anything else, and only fill out the succeeding screens with emotions or causes if the deviation is strong enough in one direction of another.

What I mean by funny here is that, as I said in therapy yesterday, I feel like if an outside, third party looked at my mood tracking they’d assume that we needed to do something get me above that midline. Whereas I think that maintaining that midline is my goal. Whenever I have an extended period (by which I mean, say, a week straight) of making “Slightly Pleasant” or even “Pleasant” on my watch, I know that one or another of the weeks that follow are going to dip the other way.

That, of course, averages out to put me squarely back at the midline overall, and I admit to some skepticism of anyone who claims that they are happy or would mark their days as “Pleasant” most of the time. I don’t look at my habitual midline as a drawback. I think I see it as part of what resilience I do have.

Me, one and a half years ago, quoting from Deb Chachra’s How Infrastructure Works:

Resilience isn’t efficient because it typically requires an upfront or ongoing investment of resources, in the form of time, money, energy, or cupboard space, in order to head off worse outcomes.

Again, though, mentally time traveling to that existential prison in western Massachusetts, were I to grade such a degradation of my autonomy on this curve, devoting these compensatory resources to convince my bodymind, somehow, that I’m on that middle path, what worse outcome, exactly, would I be heading off? The worse outcome would be here, would be my life, and I’d be grinning and bearing it for what reason? There would be no metaphorical cooler weather coming to restore my autonomy and aid my recovery. I’d just be dying, slowly, inside.

Early on in my blogging about having been diagnosed as autistic, I’d happened upon a thread I’ve followed intermittently ever since: the idea that autistic brains do not habituate. This complicates the entire question of resilience versus resignation.

(This, I believe, likely is the entire basis of autism: a sensorium dysphoria from too much getting in. It’s somewhat akin to the “intense world” theory, and monotropism is how our brains address it.)

No, I did not, in fact and inside, “get used to” the hotter summer days that kept me inside, artificially limiting my autonomy and frustrating my agency. What I realized yesterday is that these past months have completely fucking sucked, and I simply hadn’t been allowing myself to recognize just how tightly wound my entire nervous system has been the entire time.

It’s not that this realization brought an accompanying knowledge of a solution. Maybe I do have the grade the shitty times on that curve just in order to get through, and then, eventually, past them. Maybe that kind of what only can be called resignation itself is some fashion of resilience—but only when it’s known to be a transient and temporary thing.

At some point, of course, I knew, the heat would break, the days that are too warm for me to leave the house, frustratingly at increasingly lower temperatures than even just a few years ago, will give way to cooler weather and even overcast skies, and those built up tensions in body and mind can begin to unclench and dissolve.

So, yes: I’ve spent the summer grading my days on a curve. Most days were not, in fact, that middling, neutral “meh” that serves as the guiding star for my self-regulation. Yes: I grinned and bore it (without the grinning, because that’s not me), and pretended otherwise. Now I know, or think I know, that I did this, that I was capable of doing this, because eventually it would end, and I new that eventually it would end, and room could be made for recovery.

It’s just that this only works when I know, or at least when its presumptively true, that there will be a light at the end of the tunnel, even if I can’t actually see and confirm the existence of that light for months on end.

This won’t work, however, when the tunnel isn’t a tunnel, but a dead-end cave with no exit.

Reply by email • Tip $1/month