Michael Lopp lists out seven kinds of decisions ranging from the vanilla to inspired to delegated. Perhaps the one that provoked the biggest response from me is the type he calls “minimum viable”:
Your investigation into their justification reveals that they chose a decision that was not designed to be good; it was constructed to offend the fewest humans. This is not leadership — it’s fear.
I can’t tell you how many “decisions” I’ve seen in organizations which are covered with a veneer of rational-sounding company jargon. They fall apart as soon as you dig into them, leaving you with a sinking feeling as you realize the motivation is strictly political.
The saddest part is that since there is no real substance to these kinds of decisions, they quickly crumble and get replaced with newer ones of the same type. The inevitable result: the organization forgets what its real purpose is.
A few months ago, Derek Thompson wrote about one of biology’s biggest mysteries: sleep. A lot of important processes happen when we sleep, but none of these explain why sleep originated. Now, a new study reported in the Economist offers evidence that sleep exists because of volcanoes, oxygen, and the need to repair damage from electron imbalances in mitochondria.
After writing about the studies, Derek writes this beautiful take:
Achilles’ mother dipped her son in the River Styx to make him immortal, but, holding him by the foot, left his ankle dry and vulnerable. So did oxygen anoint Earth with a blessing second only to immortality — complex life, which became consciousness and intelligence. But there was a price: a kind of daily death. Rest restores our ability to burn the planet’s oxygen for energy. Why do we sleep? Because the earth once learned to breathe.
It’s a poetic thought that connects something that we take as given in our day to the origins of life on Earth creating the oxygen-rich atmosphere we breathe every day.
Most of my trip to Scotland last month was spent on the Isle of Harris, but I did get a day in Glasgow to go street shooting with Rick Lepage and Hudson Henry. My favorite photo of the day captures Rick and Hudson looking for their next shot on Sauchiehall Street while a delivery rider barrels between them. The delivery riders in Glasgow were absolutely relentless that day and we had several near misses before and after this moment.
Horsaclete, Na h-Eileanan Siar, ScotlandLeica Q3 43
Last month while hanging out with friends in Scotland, I got the chance to compare and contrast three incredible cameras side by side: The Sony RX1R III, the Leica Q3, and the Leica Q3 43. All three three sport beautiful lenses with amazing sensors. All beautifully support the process of making photographs out in the world. And, most importantly, all are really fun to use.
Of course, nobody needs to own one of these cameras. There are less expensive alternatives that will allow you to express your creative vision. Some of us, however, are undeniably attracted to them because of their size and the way they let us pursue our vision in the world. They let you move through the world in a different way than larger cameras do. When you’re in the creative mindset, sometimes that matters.
Since everyone in our group was more than happy to swap and share, I was able to spend quality time with each of these cameras while out in the Scottish landscape to compare and contrast them, and to find out which one I wanted to spend more time with.
To cut to the chase, I think the primary reason to get the RX1R III is if ultimate portability is non-negotiable. It’s the perfect camera for traveling as light as possible when every single milliliter of space and gram of mass matters. It’s the iPhone Air of the bunch, if you will. It’s an ideal camera for travelling on business when you really want everything in a single bag and the camera is just coming along for the ride.
Outside of that very narrow niche, the Leica Q3 wins. There is a magic to the Q3 that the Sony just can’t match. The lenses are legendary. And, honestly, they’ll travel almost as well on a trip where you want to travel light unless you really have serious size and weight constraints.
Now that I’ve given my conclusion, if you’re interested in reading on, I’ll tell you how I came to it.
Rodel, Na h-Eileanan Siar, ScotlandLeica Q3 43
Backstory
Smaller cameras like this have always been attractive to me. When I was a kid, my grandparents had a Leica M and a Leica CL, which they taught me how to use and shoot film with. As a geeky adult, however, I quickly picked up the first digital cameras and never looked back. I made do with a variety of small digital cameras, but nothing really fit until the original Sony RX1 was released in 2012.
With a Zeiss Sonnar T* 35mm f/2 lens fused to a good full-frame sensor and not much more, the RX1 was a revelation. Small enough to fit in a medium size jacket pocket, it truly was the best camera you could have with you most of the time. The compromise was that it had shit for battery life and didn’t have an integrated viewfinder, but the image quality and ultimate portability was worth it.
The first digital Leica that really caught my attention was the Leica Q, released in 2015. It was a pretty amazing camera for the time, and the Summilux 28mm f/1.7 ASPH lens was truly excellent, if a bit wide for my taste. It made beautiful images — even people who didn’t know about cameras noticed the difference in the shots I made with it. But, it didn’t ever travel as comfortably with me as the RX1 did.
When the Leica Q3 was released in 2023, it replaced the good enough sensor of its predecessors with a truly great one. I still didn’t love the 28mm focal length, but I figured the Sony RX1 line was done for and there wouldn’t be another one, so I purchased one. And discovered it had a minor-sounding feature that was great for traveling: being able to charge over USB-C and not carry a separate charging brick. Minor sounding, but huge for portability.
In 2024, Leica released the Q3 43 with a APO-Summicron 43mm f/2 lens. Dammit! I was instantly torn. I knew I’d like the focal length more, but I wrestled with the idea of selling my Q3 off to fund it. Other things took my attention and I pushed off the thought for later. Then in a surprise announcement this year, Sony announced an updated RX1R III with a state-of-the-art sensor, the same Zeiss 35/2 that I’ve loved over the years, and updated electronics, including essentially the same sensor in the Leica Q3.
What a quandary! RX1R III or Leica Q3 43? I knew that I probably wanted to move on from my 28mm Leica Q3 to one of these, but which one? That’s where I got lucky in joining Hudson Henry and Rick Lepage in Scotland after they wrapped up two photography workshops to go scouting. Between everyone in the group, we had all three cameras to compare.
Luskentyre, Na h-Eileanan Siar, ScotlandSony RX1R III
Luskentyre, Na h-Eileanan Siar, ScotlandSony RX1R III
Where the RX1R III leads
I’ll start with the RX1R III and the things it has over the Leica Q3 that matter: size and volume, autofocus, and the way its digital crop works.
Comparing height, width, and depth numbers doesn’t communicate how tiny the Sony is in comparison. It’s the volume and weight of the camera that really sets it apart. It’s something like half the volume of the Q3. It’s smaller than the APS-C Fujifilm X100VI! And it weighs under 500g, compared to almost 800g for the Leica Qs. You can simply fit it in with more things in your backpack, and you can even sneak it into some jacket pockets.
The RX1R III also has the better autofocus. While Leica’s autofocus has consistently improved over the years and is now firmly in the good category, the RX1 has Sony’s latest preternaturally great autofocus. It’s faster and so more confident than Leica’s system. You never have to think of it. The Leica autofocus isn’t bad, but Sony’s is on a whole different level.
Finally, the way that Sony’s digital crop works is better. In a stylistic nod to Leica’s rangefinder history, the Leica Q3 puts in crop lines while you see the whole frame of the sensor. It’s sentimental and cute and it‘s representative that you’re going to get the full 61 megapixels in a raw file. The RX1R, however, just zooms the image you see in the viewfinder or on the back screen. When you’re shooting at a 50mm equivalent crop, you see that crop. What you see is what you get.
Rodel, Na h-Eileanan Siar, ScotlandLeica Q3 43
Urgha, Na h-Eileanan Siar, ScotlandLeica Q3
Where the Q3 pulls ahead
Compared with the RX1R III, the Leica Qs have noticeably better lenses, a simpler and more intuitive user interface, a better EVF, image stabilization, weather-sealing, a flip-screen, and integrate better with an iPhone or Android device.
As good as the RX1’s Zeiss 35/2 is — and it does deliver the goods every time — the two Leica Q lenses are just that much better. Not only do they deliver the goods, they do so in a way that’s almost otherworldly. I was frequently surprised with the results when reviewing images, sometimes to the point of saying expletives under my breath. The images, especially from the 43mm lens, often give me the same feeling that I used to get reviewing chromes on a light-table with a loupe compared to looking at a print made from a negative.
The user interface of the Q3s, both physical controls and software, feels like they were designed by photographers for photographers. Intuitive and natural, everything does what you expect it to do and doesn’t require futzing around at all. While the RX1 has Sony’s latest and much-improved menu system, it’s still more complicated and what you want is often a bit buried compared to the Leica interface. As well, the physical controls on the Sony are cramped in comparison to the Leica. Sony could follow Leica’s lead here and simplify more.
The RX1’s EVF is… perfectly usable. It’s small with a small viewing area, but that’s understandable because the camera is small. I’m not sure what more can be done there. But, not only is the Leica Q3s EVF bigger, it has double the resolution and is much more comfortable to use. It’s like when we went from regular resolution displays to Retina displays on the iPhone. It’s something that you didn’t know you wanted till you have it and then you really want it.
Moving on to image stabilization, the Q3’s system is really good. The RX1 simply doesn’t have image stabilization. A decade ago, the charm of having the smallest full-frame camera in the world more than made up for the lack of image stabilization. In 2025, it’s a bit of a handicap. The same goes for weather sealing. I never had a problem with my RX1 in inclement weather, but I wasn’t ever as cavalier with it as I was with other cameras. It’s a complete miss that Sony didn’t put better weather sealing into the RX1, especially since the gaskets wouldn’t weigh anything or take up any more space.
As far as the flip screen, it’s baffling that this iteration of the RX1 doesn’t have one, especially since the previous mark II iteration did. Not having it in the current version feels like a serious regression, especially since the EVF is so cramped. I wasn’t sure this was a relevant factor until shooting the cameras side by side and learning how much I did like using the flip screen while walking the streets of Glasgow or hiking on dunes on the Isle of Harris.
Finally, and this also didn’t really sound like much of a deal until using the cameras side by side, Leica’s iOS app is really good and integrates with the Q3 perfectly. By comparison, the Sony app gets the job done if you want to review or pull images, but it’s rather clunky. It only connects to the camera when the camera is powered on, while the Leica app can wake your camera up in the bag to review images. And the camera can always connect to the app while your phone is in your pocket to grab a GPS update to tag images with the location of where you are.
Scalpay, Na h-Eileanan Siar, ScotlandLeica Q3 43
Scalpay, Na h-Eileanan Siar, ScotlandLeica Q3 43
Where I landed
If any of these cameras existed in isolation from each other, I’d be hard pressed to complain about anything, other than the price tag. If the Leica Q3s didn’t exist, the RX1 would be a no-brainer for me. It that was the camera I had bounced around Scotland with mucking through the bogs finding odd stuff to photograph, I would have been more than happy with it.
It’s in comparison where the differences stand out. And because of those price tags, what we’re looking for really comes down to the joy factor. You buy cameras like this to make images, which includes both the process of making the image as well as the images themselves. You want a camera like this to spark joy when carrying it, using it, and looking at the results.
So, I come back to my conclusion: The biggest thing the RX1R III gives you is ultimate portability. It can go more places more often with you. But for me, the Leica Q3 — in particular the Q3 43 — pulls ahead.
When I first picked up the Q3 43 and put it up to my eye and framed a shot, I muttered “mother fucker” under my breath. And then when I saw the shot, I said the same thing again, quite a bit more loudly. The frame is perfect for me. I had taken the claim that the 43mm lens matches normal vision with a bit of salt because our vision isn’t rectangular and the periphery is large, but I think Leica is onto something with this focal length. It really works for me.
That said, in our group we were evenly split on our preference for the 28mm or 43mm lens. Half of us really preferred the 28mm frame. The other half preferred the 43mm. And, of course, we all made the observation that sometimes you really do want a camera with a good zoom on it so that you can pick your frame in the field, even if these Leica lenses have a pop that few other lenses can even hope to approach.
I’ll wrap up by saying again that nobody needs any of these cameras. In fact, you absolutely shouldn’t rush out and get one of these cameras unless you really know what you’re getting into and aren’t breaking the bank doing it. But, if you are in a position where you are considering these cameras, I hope this helps. As for me, I’m going to go figure out what to sell as I couldn’t quite bring myself to nick my friend’s Q3 43 before I left Scotland.
Horsaclete, Na h-Eileanan Siar, ScotlandLeica Q3 43
You’d never imagine that you’d find a speciality coffee shop on the empty road between Tarbert and Luskentyre on the Isle of Harris. And certainly not in a vintage caravan housed alongside the wild moorland. But there is and it’s called Isle Coffee Harris. It’s reportedly quite good, and some days you can even get a cinnamon swirl with your cortado. Unfortunately, we didn’t get to sample it ourselves, as we drove by either too early or too late.
I’ve been really hoping that Claude Code would shift to support AGENTS.md as a standard for providing default instructions about a project. Sadly, that hasn’t happened yet. I’ve been hacking around this by symlinking CLAUDE.md to AGENTS.md. It works, but I find it to be subpar and maybe even a big gross.
Just north of Pitlochry in the Scottish Highlands is the Killiecrankie river gorge, carved by the River Garry. It’s crossed by Garry Bridge which stands 20 meters over the water, and has a permanent bungee jump platform underneath operated by Highland Fling, if that’s your thing.
The area is the site of the 1689 Battle of Killiecrankie, one of the bloodiest battles in Scottish history — a history full of bloody battles. The Jacobites won the battle but lost their commander, Viscount Dundee, which led to the collapse of that particular uprising.
I made this photo during a quick stop on a drive from Glasgow to Ullapool on our way to the Isle of Harris. We didn’t spend much time, but I would have loved to explore the gorge with its colors.
Take a flight to Glasgow. Join a friend who has a friend with a car. Drive north through Inverness to Ullapool to catch the ferry to Stornoway. Once you’re across the water, drive to Tarbert and then follow a set of semi-complete directions past the post office. Keep going up by the recycling center and then look for a really small sign next to a turn off to a gravel track road. Follow the road down to a little white house by the edge of slate-blue water that’s so much clearer than you’d expect. That’s how I got to Lingerbay on the Isle of Harris in Scotland.
It’s quiet here. At 58ºN, it’s not the furthest north I’ve ever been, but it’s the furthest north I’ve been in Great Britain. Things move slow. It’s chilly, and moody, but somehow lovely all the same. The weather changes its mind a lot. At times, it feels grey, and then you notice turquoise in the water and golden in the moorland. There’s very slow 4G service which keeps us just connected enough to the world, and no more.
We’re here to make pictures, talk photography as much as we can, and have a bit of whiskey while we’re at it. Politics keeps coming up, and we indulge it for a bit, and then shove the topic aside in favor of talk of composition, seeing, and being somewhere for a bit without the need to rush.
Despite being a mission-critical part of Ruby since it was created by Chad Fowler, Rich Kilmer, and others at RubyConf in 2003, RubyGems has remained outside the core Ruby community. At least until now.
To provide the community with long-term stability and continuity, the Ruby core team, led by Matz, has decided to assume stewardship of these projects from Ruby Central. We will continue their development in close collaboration with Ruby Central and the broader community.
Oh thank god. This is probably twenty years overdue.
For a while after its creation, RubyGems was loosely supported by the community. Then, Rails-hosting provider Engine Yard stepped in and provided support until 2015 or so. At that point, André Arko and several others formed Ruby Together to serve as a vehicle for sponsorship of the work on RubyGems and Bundler. But, Ruby Together had challenges with securing consistent funding, and ended up merging with Ruby Central in 2022.
That didn’t help as much as it should have and things have been in an uncomfortable state since then. During my entire tenure at Shopify, I worked closely with the Ruby and Rails teams there and we were incredibly concerned about the security of the Ruby ecosystem software supply chain and I know that concern has remained since I left Shopify.
What a mess. I personally believe that everyone was acting with the best of intent from their point of view under a lot of pressure. I can’t help but think how much better this would have gone if everyone had sat down first over a meal and a beverage to talk through how to accomplish this. In the end, however, I think putting the RubyGems repository under the main Ruby organization is the right move. The next question is what happens with the RubyGems service. It’s been a bulletproof service for a long time and I hope it remains so for a long time.
Nine months ago, Chris Anderson kicked off a Willy Wonka-esque search for renewing the leadership of TED. At the time, I have to admit that I was pretty skeptical. Maybe it was a fear of the unknown. More likely, a fear that whoever Chris sold TED to would be someone or something that would take TED into a bad direction.
I was also a bit intrigued because this opened up a possibility space for an organization that I invested a lot of sweat equity into from 2009-2016 as a main stage photographer. Lately, TED has felt like it’s stagnating and playing it too safe in its current formula. A little too comfortable with the glossy talks. A desire to be brave with controversial talks, but a bit too cautious of offending someone on either side of the political spectrum.
Therefore, I was nervous — as were many people — to listen in to yesterday’s announcement of the result of the process. I was very relieved to hear that despite the chance to sell out or to transform TED into a for-profit endeavor, Chris doubled-down on the idea that TED meant something and should to be preserved. In the blog post of the announcement, Chris says:
But for me — and for everyone on TED’s leadership team — the answers weren’t just about capital or scale. They were about stewardship, values, and a shared belief in giving ideas away, trusting community, preserving independence, and amplifying human possibility.
When Chris then announced that Sal Khan would be taking the position of Vision Steward, I was initially surprised, and then a mixture of relieved and intrigued. Vision Steward is certainly an interesting title, but it’s clear that Sal is Chris’ replacement as the person who curates and shapes TED’s direction. Sal has done amazing work with Khan Academy and I think he can help TED a lot in this new role and bring a very fresh perspective.
I was also excited to see Logan’s appointment to CEO. I first met Logan at the very first TED I photographed in Long Beach in 2009, and I’ve been a fan of hers since the day I met her. She was so full of potential then and has certainly become a powerhouse now. Her recent work at both the Obama Foundation and with TED Countdown has been amazing to see. And while these are fairly left-leaning credentials these days, she also worked for almost three years at Palantir, which is quite the counter balance.
Look, every change is risky. And I know that there are people in the wider TED community that were underwhelmed by the news. Maybe they were seriously hoping that a firebrand billionaire like Elon Musk would do something like merge TED with what’s left of Twitter. Or maybe they were hoping for a buyer that would move it away from the United States and the current political turbulence there to a safe haven where it couldn’t be repressed into insignificance. Maybe they’re right, and this won’t be enough to revitalize TED. Maybe TED is from a different moment in time that can’t make the jump to the new normal.
I don’t know. We’ll have to see. What I do know is that I have a huge amount of trust in the worldview of both Sal and Logan. If they are truly unshackled and ambitious — and maybe even are more than a bit audacious in really tweaking the formula and running further into the controversial and meaty topics that we need to make progress on both in the United States and in the world at large — and they succeed in reinventing TED, then it’ll be a TED that I will be excited to see in the world again. I hope that’s a TED that can help pull us back from the toxic negative-sum discourse we are drowning in and into a positive-sum discourse that will help us face the future that’s quickly coming, as uncomfortable as it is.
A few months ago, I found myself in the Bay Area for a quick overnight and hotel rooms in San Francisco were outrageously priced. So, I stayed across the bay in a room with a front-row view of Oakland’s City Hall with its distinctive clock tower. Completed in 1914, it’s a beautiful example of American Beaux-Arts architecture and stands in contrast to the more modern buildings around it.
I was jet-lagged and up before dawn, so I was able to enjoy the view from my window as the sun rose over the East Bay hills and bathed the scene in a hazy golden light. This is one of the first frames I captured with the recently released and somewhat controversial Sony RX1R III. I don’t have a judgment yet myself, other than it’s really nice to shoot with that lovely 35mm lens again. It’s a focal length that just agrees with the way I see the world.
Slop online — low-quality, generic, bland, shallow, unreliable, attention-seeking, mass-produced content — isn’t new. Humans have been making it for as long as the internet has existed, because attention is the only valuable resource online. But AI amplifies this problem by orders of magnitude, and Kurzgesagt has made a video that dives into why this isn’t great.
If current trends continue, it’s not that far fetched that cheap slop content, stuff just good enough, will soak up the majority of human attention. It could make us dumber, less informed, our attention spans even worse, increase political divides and make us neglect real human interaction. If AI eats the majority of the attention pie, channels like ours will become unfeasible or forced to downsize or use AI themselves to be able to compete. We don’t want to play this game.
That doesn’t mean that they want to be Luddites and not use AI at all. In the last-act of their narrative, they promise that they will use AI as a tool in their toolbox like the align tool in Adobe Illustrator or as a better Google search, but the creativity and integrity in their video is still theirs. They promise that “Kurzgesagt is made by humans, for humans, and will remain that way.”
Made by humans, for humans. I love this idea. It matches how I’ve been using AI. When I write these posts, for example, I ask a LLM to rate and review my posts and make suggestions, but it’s me doing the writing. Em-dashes and all.
Maybe LLMs believe a seahorse emoji exists because so many humans in the training data do. Or maybe it’s a convergent belief - given how many other aquatic animals are in Unicode, it’s reasonable for both humans and LLMs to assume (generalize, even) that such a delightful animal is as well. A seahorse emoji was even formally proposed at one point, but was rejected in 2018.
Sure enough, when I prompt GPT-5 asking for a seahorse emoji, I get this reply:
Hilarious. This is the 2025 version of the “How many r’s are in strawberry” prompt that would trip up most models until recently. It shows how LLMs are pattern learning and matching machines and can generate what seems like it ought to exist rather than what actually exists — similar to the Mandela effect experienced by large group of humans that share a false memory.
Frank Chimero’s latest essay applies The Market for Lemons — a concept from the 1970s about information asymmetry in markets — to explain why so much of the internet feels broken right now. You can see this in action on Amazon these days with generic random letter brands like PCYTECH and HMIYA showing up everywhere. Sellers are putting more efforts into gaming the system than actually making a great product.
What makes the Market for Lemons concept so appealing (and what differentiates it in my mind from enshittification) is that everyone can be acting reasonably, pursuing their own interests, and things still get worse for everyone. No one has to be evil or stupid: the platform does what’s profitable, sellers do what works, buyers try to make smart decisions, and yet the whole system degrades into something nobody actually wants.
This pattern shows up everywhere and Frank’s essay really hits home for me when he applies the thought to the recruiting process. Companies have a hard time differentiating between great candidates and mediocre ones. And candidates have incomplete information about how the company works or what the product roadmap is. Gaming the system becomes essential, making the only reliable signals reputation and connections. Everything else is suspect even though most people aren’t operating from a position of malice.
Our hypothesis was that LLMs are largely conditioned to mirror human behavior because of how LLMS have traditionally been trained. If we present journaling and social media as natural affordances, agents would engage with them organically rather than mechanically following prescribed workflows, much like a human would.
Not wanting to use a real social network, they created their own and then encouraged their agents to use it. They stood back and watched as agents developed sophisticated behaviors and build on what other agents had written. For example, when the agents got stuck, they could look at what other agents had posted about similar issues and build on those ideas.
Is this emergent behavior baked into models thanks to being trained on the Internet? Or is this something else where agents organically are adopting social behaviors to improve their own work? It’s really not clear. My bet is that, like sycophancy, it’s a quality that’s comes training masquerading rather than true emergence. But, it’d be really amazing if there was something deeper about how collaboration improves outcomes.
I really don’t like the narrative that we have 18 months until AI leaves us in the dust. Derek Thompson, co-author of Abundance with Ezra Klein, has found himself in the same place. Instead of thinking we’re doomed by the machines, he thinks the real crisis will be people voluntarily degrading their own cognitive abilities by outsourcing their minds and letting their reading and writing skills devolve. In his recent post, The End of Thinking, he writes:
One might imagine, then, that the potential arrival of all-knowing AI would serve as a galvanizing threat, like a Sputnik moment for our collective capacity to think deeply. Instead, I fear we’re preparing for the alleged arrival of a super-brain by lobotomizing ourselves, slinking away into a state of incuriosity marked by less reading, less writing, and less thinking.
It’s a scary thought. Literacy is a cornerstone of how we summoned the modern world into existence. It lets us transmit ideas across space and time in ways far beyond what oral traditions could do. To let it atrophy now, in the age of AI, would be to fulfill Neil Postman’s warning in Amusing Ourselves to Death — handing over our best selves not to tyrants, but to distraction. Keeping it alive must be one of our highest priorities.
Jerry Neumann takes a deep look at the investments happening in the AI space and concludes that AI will not make you rich, at least if you look at it like previous innovations:
The way to invest in AI is to think through the implications of knowledge workers becoming more efficient, to imagine what markets this efficiency unlocks, and to invest in those. For decades, the way to make money was to bet on what the new thing was. Now, you have to bet on the opportunities it opens up.
Like the web in the late 90s, we’re definitely in the frenzy zone. There‘s almost certainly a bust of some sort coming. But the really interesting stuff is what will come next. The question I’m asking myself is: What can you do to be ready for that? That doesn’t just include learning how to use the LLMs well. It extends to looking for the second- and third-order opportunities that are just over the horizon.
Claude’s memory system has two fundamental characteristics. First, it starts every conversation with a blank slate, without any preloaded user profiles or conversation history. Memory only activates when you explicitly invoke it. Second, Claude recalls by only referring to your raw conversation history. There are no AI-generated summaries or compressed profiles—just real-time searches through your actual past chats.
These searches are accomplished through the conversation_search and recent_chats tools. You can prompt Claude to use these by writing something like, “Can you tell me what we talked about in our last 4 conversations?” You can even get Claude to give you information about these tools by asking, “What tools do you have available for dealing with chat history?”
This is very different than how ChatGPT’s memory works using a combination of a set of user-controlled memories and a dense generated summary generated from conversation histories that aren’t user visible.
I’m not sure which approach I like better. I think I prefer how Claude gives more explicit control over what goes into the context window as compared to not knowing what is in the generated summaries of previous conversations that are updated on an opaque timeframe.
The differences stood out immediately. Noise cancellation is far stronger, making it easier to hear details even in a loud room. The bass is tighter and the highs are brighter. The new foam ear tips feel secure and comfortable, while transparency mode is so natural it feels like nothing is in your ears.
It’s a ringing endorsement and sounds like a great update to the product. The AirPods have been a hands-down awesome product for Apple, and I’m really looking forward to getting my hands on this latest version.
Right on schedule, this year’s Apple September announcements arrived in yesterday’s “Awe Dropping” event, bringing updates to iPhones, AirPods, and Apple Watches.
Everything announced seems like a solid update. It’s great to see AirPods Pro 3 get next-gen noise cancellation, better ear fit with new foam-infused ear tips in five sizes, and heart rate monitoring, and the satellite connectivity in the Apple Watch Ultra 3 is a welcome addition to an adventure watch.
As to the iPhones, the new iPhone Air is very cool looking. I find it a really tempting design and I’m sure it will feel really great in hand, and it’s great that bumpers are back, though I want to hear how battery life actually works out in the real world. Of course, I’m camera obsessed enough that I’ll probably skip the Air and upgrade instead to the iPhone 17 Pro — especially now that both sizes feature the same camera specs. And, I’m actually glad that they’ve taken the Pro back to aluminum for heat dissipation. Sure, titanium is nice, but my current iPhone 15 Pro definitely feels the heat in the summer.
There’s always something that only gets mentioned in the supporting material that I get excited about which isn’t ever shown in the events. This time, it’s that Final Cut Camera 2.0 brings open gate recording as well as the ProRes RAW and genlock support that were mentioned. Of course, most people won’t care about being able to do open gate recording — which is where video is captured using the entire 4:3 sensor so that you can choose your crop later — but for those that do, it’s really exciting. It’s the kind of detail I wish I had on some of my big Sony cameras.
One other thing that wasn’t mentioned at the event is that release candidates were released for all the operating systems. I’ve sat out of most of the beta cycle this time, but now that the RCs are out, I’m diving fully in across all my devices. And… I kind of like a lot of what they did with Liquid Glass on iOS in the end. Certainly, I like it more than I thought I would after seeing the demos during WWDC. On the other hand, the big rounded corners on windows in macOS are going to take getting used to.
The event itself was like all of Apple’s events of late, shot on iPhone, tightly edited, and super-produced with swoopy aerial transitions. It’s nice to see them using locations outside of Apple Park, including Apple’s signature stores in San Francisco, Miami, Chicago, and Brooklyn. They have built up a really amazing portfolio of architecture and it’s cool to see them showing it off. On the other hand, it’d be nice to see Apple break a bit out of their current presentation formula. They’ve perfected it to the point where it’s missing something. A je ne sais quoi. They lack emotion and are too clinical now.
But even with that critique, this event has become part of the end-of-summer and back-to-school feel of September. And this year, everything I’m interested in gets a solid update. I have quite a few friends that work at Apple, and I’m always happy to see the things they work on come out.
