Categories
AI Blogs/Weblogs Living Menlo Park

The Foothills

It was later in his illness. Someone had set up a folding table in the garage and Chris was sitting at it in a folding chair, working through a stack of photographs. Signing them, one by one, telling me the story inside each one as it came up โ€” where heโ€™d been, what was happening just outside the frame, what heโ€™d seen in the viewfinder that made him press the shutter at that exact moment and not a half second later. The garage was quiet. Outside, Menlo Park was doing whatever Menlo Park does on an ordinary afternoon. In here, a man was accounting for his life in pictures and I was standing there holding a camera, not quite sure what I was witnessing.

I made a photograph of him.

Itโ€™s at the top of his Wikipedia entry now. Thatโ€™s how the world knows his face โ€” a picture I made of him making sense of his pictures, in a folding chair, near the end. I donโ€™t know what to do with that except carry it.


Chris Gulker had been a photographer long before he was anything else. Staff photographer at the Los Angeles Herald-Examiner. Twice nominated for a Pulitzer. Published in Time, Newsweek, Rolling Stone. He had the eye first. Everything else โ€” the virtual newsrooms, the blogrolls, the hacked-together color systems that dragged newspapers into the digital age โ€” all of it came from the same instinct: look carefully, see whatโ€™s actually there, build toward what you see.

When I first met him he had just gotten a Leica M8. He talked about it the way he talked about everything he loved, which is to say with specificity and without apology.

He had driven an Audi TT. He had a Leica M8. He was not a man who made concessions to the ordinary.

He had glioblastoma. Diagnosed in 2006. Surgery, radiation, the whole negotiation with a disease that doesnโ€™t actually negotiate. He knew the terms and he kept going โ€” kept shooting, kept writing at gulker.com, kept thinking out loud about what was coming next, as if the tumor were an inconvenience and the future were the point.

He walked when he could walk. He talked when he could talk.

He died in October 2010. He was fifty-nine.


Twice a week in those last two years Iโ€™d put Lily in the car and drive over to his house. Lily was small and opinionated and she understood the trip as hers. Weโ€™d pick Chris up after breakfast, when the morning was still cool, and do the loop โ€” one mile, flat, because flat was what worked. Then weโ€™d come back to find Linda moving through the house, Chrisโ€™s wife of nearly thirty years, the still point of everything that was happening to them. Sometimes sheโ€™d join us and the conversation would open into something more alive, the kind of talk where someone says something offhand and suddenly everyone is leaning forward.

One of those mornings the three of us decided to start a local blog for Menlo Park. Linda would write and edit. Chris would shoot. We called it InMenlo.com.

When Linda wrote Chrisโ€™s obituary, thatโ€™s where she published it.

People talk about spending time with the dying as a kind of grace extended downward. It wasnโ€™t like that. Those mornings were a gift โ€” the ideas, the talk, the way Chris described what was coming as if he could already see it clearly from wherever he was standing. I left those visits more alive than I arrived. Thatโ€™s the debt I carry. Not grief exactly, though thereโ€™s grief. More like an obligation to keep paying attention to the future he spent his life building toward.


Last month a man named Demis Hassabis closed a two-hour technology showcase in Mountain View โ€” twenty minutes from where Chris and I used to walk โ€” and said seven words I havenโ€™t been able to put down since: We are at the foothills of the singularity. The audience applauded. Then everyone went home.

I keep thinking Chris would have had something to say about that.

Not the singularity part, necessarily โ€” that word carries a slightly rapturous charge, too certain of its own prophecy. But the foothills part. The careful humility of it. The acknowledgment that what we can see from here โ€” AI systems autonomously building operating systems, models that predicted a hurricaneโ€™s landfall and saved lives โ€” all of it is still just approach terrain. The mountain is what comes after.

Chris spent his whole career in the foothills of things. Slightly ahead of the moment, always building infrastructure for a future that hadnโ€™t arrived yet, always explaining to people who werenโ€™t sure they wanted to know. He pioneered the blogroll. Built one of the first online newspapers. Hacked color into the San Francisco Examiner with Macintoshes and ingenuity when the system said it couldnโ€™t be done. He was the wrong man for the present tense. He belonged to the next sentence.

He had the photographerโ€™s instinct underneath all of it โ€” the knowledge that you have to look carefully, that the light is always changing, that if you wait too long the moment is gone. He put the Leica to his eye and he saw. He put his hands on a keyboard and he built what he saw toward.


Lily is gone now too. She outlasted Chris, which felt right โ€” she was stubborn and she loved the route.

I still think about those mornings. The cool air, the flat mile, Lily pulling us both forward. The way the real conversation started when we got back. The way Linda might appear and the whole thing would open into something none of us had planned. The way Chris talked about what was coming โ€” not as speculation but as something he could already see, the way a photographer sees the shot before he raises the camera.

He always knew something was coming. He had a gift for the future tense Iโ€™ve never quite encountered in anyone else โ€” and a photographerโ€™s understanding that the future, like light, doesnโ€™t wait.

I wonder what heโ€™d make of the foothills. I think heโ€™d already have the Leica out. And I know weโ€™d still be talking about it.

Categories
AI

The Jagged Mind

There is a peculiar kind of genius that has always made us uneasy โ€” the savant who can calculate the day of the week for any date in history but cannot tie his own shoes. We admire the capability. We are troubled by the gap.

Demis Hassabis, speaking at this weekโ€™s India AI Impact Summit in Delhi, gave that unease a name. He called todayโ€™s most powerful AI systems โ€œjagged intelligences.โ€

It is a phrase worth sitting with.

A jagged intelligence can win a gold medal at the International Mathematics Olympiad โ€” solving problems that would humble most PhD mathematicians โ€” and then, in the very next breath, stumble on elementary arithmetic if the question is phrased in an unfamiliar way.

The peaks are extraordinary. The valleys are bewildering. And crucially, you never quite know which terrain youโ€™re standing on.

Hassabis identified three specific gaps between where we are and what he called โ€œa kind of general intelligence.โ€

The first is continual learning โ€” todayโ€™s models are trained, then frozen. They are, in a sense, educated and then released into a world they can no longer learn from.

The second is long-term planning. Current systems can reason tactically, but they lack the capacity to hold a coherent thread of intention across months or years the way a human architect, scientist, or entrepreneur does.

The third โ€” and perhaps the most philosophically interesting โ€” is that jaggedness itself: the wild inconsistency that makes todayโ€™s AI feel more like a force of nature than a reliable mind.

โ€œA true general intelligence system shouldnโ€™t have that kind of jaggedness.โ€

What strikes me about Hassabisโ€™s framing is how it reorients the conversation.

We have spent years debating whether AI is โ€œintelligent.โ€ His point is more subtle: intelligence without consistency is not yet wisdom. A system that is brilliant and brittle in equal measure is something genuinely new in the world โ€” not human, not the robots of science fiction, but a third thing we donโ€™t yet have good language for.

The road from jagged to coherent is, I suspect, the central engineering and philosophical challenge of the next decade.

Continual learning means systems that grow with us. Long-term planning means systems that can be trusted with consequential goals. Consistency means systems whose judgment we can actually rely on.

Until then, we are working with something that resembles a prodigy โ€” dazzling, occasionally humbling, and not yet quite whole.

Questions to Consider

  1. The Consistency Problem: If you knew an AI system could solve a problem brilliantly 90% of the time but fail unpredictably the other 10%, how would that change the decisions youโ€™d trust it to make?
  2. Frozen in Time: What does it mean that the systems we rely on most are, at their core, educated in the past and unable to learn from the present? What human analog does that bring to mind?
  3. Jagged vs. General: Hassabis draws a line between โ€œjagged intelligenceโ€ and โ€œgeneral intelligence.โ€ Do you think general intelligence is the right destination โ€” or is there value in systems that are deeply specialized, even if inconsistent?
  4. The Savant Question: Weโ€™ve always had a complicated relationship with uneven genius in humans. Does the โ€œjagged AIโ€ problem feel categorically different to you, or just a new version of an old puzzle?