sangsara.net

Tag: Artificial Intelligence

  • Week 21.25

    Week 21.25

    • This week’s installment is update #256! That’s a big deal for fans of computationally significant numbers.
    • Oddly, just as Michael blogged that his family was down with gastroenteritis last week, a similar bug hit our household. Kim got a bad case of what we thought was food poisoning, and then, of course, I came down with it two days later. It’s a weird one: the stomach trouble comes with headaches and fatigue. Fearing that it was contagious, plus being all tired out, I had to skip one meetup and a wedding dinner over the weekend.
    • So not that much happened with me this week, but I suppose we can talk about the mess out there?
    • Google held their annual I/O event and showed off their latest AI achievements. Tl;dr, some of this stuff is just gross. From a technical standpoint, yes it’s remarkable that pretty realistic video (with sound) can be generated from text, and Google can now use your personal context from documents and emails to help with tasks — similar to what Apple promised (but has yet to deliver), with the important distinction that one centers privacy and on-device computation while the other will do it on their servers. I don’t know if I trust Google to let an AI crawl all my documents, and for that reason, minimized my exposure to Google years ago.
    • I remember when Gmail first came out and pioneered showing contextual ads next to your emails. There was an uproar, and the company had to calm people by saying ‘no one is reading your emails, it’s just automated keyword matching’. Well, with LLMs, it’s much, much worse. No human would be able to go through all your messages, photos, location data, and search history to piece together an invasive psychological profile about your vulnerabilities, and make it actionable for advertisers. Trust that an AI will. Just look at how a version of Claude 4 Opus in testing tried to blackmail an Anthropic engineer over an affair it believed they were having.
    • Beyond the business model of Google’s AI products, it’s their designed intent that feels particularly bleak. One example they proudly demoed: a friend emails asking for holiday recommendations at a place you’ve visited. Instead of writing them a thoughtful reply, you let Google AI scour your photos, emails, documents, and receipts to auto-generate a message, using words chosen to sound like you. Sundar Pichai even had the nerve to say, “With personal smart replies, I can be a better friend”. With friends like these, don’t even bother writing. Just ask Google directly and it can snoop the inboxes of a billion other customers to give you the statistically “best” itinerary.
    • And this is where a company’s lack of imagination and care makes itself plainly apparent. Instead of designing an AI system that writes replies in your place, they could have made one that recaps your holiday with a little presentation showing you where you went, what you did, and what you enjoyed most. Then, memory suitably refreshed, you could sit down and write your friend a reply that shows you actually give a shit about them, and both of your lives would be richer for it. I’m beginning to think that Apple, by failing to ship their AI features on time, might be saving us from a future I don’t want to live in. Maybe they’ll never ship. Maybe that was the plan all along.
    • Then a man who many would expect to know better — who sits on stages and professes the importance of values in technology, and is arguably the most famous designer of this century — announced a new venture (also called io) with Sam Altman and OpenAI. And the vibes, my friends, were off. The 9-minute launch video came across as a thin PR exercise to polish Altman’s spotty public image and reassure OpenAI’s investors. It struck an uncharacteristically self-congratulatory tone for the usually humble Ive while announcing, essentially, nothing. Sterling Crispin tweeted a biting Marxist read of the video in the style of Slavoj Zizek. Many saw it as Altman angling for a Steve Jobs comparison, with critics pointing out that he’s not enough of a product person to be a true partner (and necessary editor) to Ive. In any case, I have no doubt that the io team will deliver some beautifully designed hardware. But I fear even they can’t summon enough thoughtfulness or optimism to divert AI from its current trajectory, or prevent the cultural and societal wreckage it’s likely to create.
    • Speaking of nice devices, one of my fondest gaming memories involves my last year at university, when I switched from my PC to a Mac that was relatively useless for playing games. I couldn’t imagine not having any games (this was before smartphones, mind you), so I bought myself a Game Boy Advance SP — the pinnacle of the series, in my opinion. It’s hard to imagine a more perfect form factor for a handheld console. The clamshell design protects the screen; the vertical layout keeps your arms and hands close together, even when squeezed into an economy flight seat; the screen was sharp and self-lit, which was not a given in those days. I would lie in bed in the dark on cold nights and play Final Fantasy I until I was bored to sleep. To this day, I still fall asleep during boring turn-based battles — one of the reasons why I want to try Clair Obscur Expedition 33, since it blends turns with real-time actions. Pretty sure I traded in my GBA SP for store credit to buy the Nintendo DS (DS Phat) when it came out later that year, and I regret it.
    • Anbernic, the China-based maker of retro emulation handhelds, released a clone of the GBA SP last year, the wonderfully named RG35XXSP. It runs a Linux-based operating system and may or may not come preloaded with thousands of classic gaming ROMs from the NES to the PS1 — you should only play the ones you actually own, of course. I didn’t get one because Anbernic doesn’t have the best rep when it comes to build quality, and I wasn’t sure they’d get it right the first time. Fast forward to today, and the new RG34XXSP (yes, it went back a model number) looks to be the one to get, even over the competing Miyoo Flip V2. DYOR (do your own research) though! They made it slightly smaller, dampened the button clicks, added two analog thumbsticks for wider compatibility, and came up with some better colorways. Now that it finally sounds more like a product than a prototype, mine’s on its way in the mail, and I can’t wait to relive some old games while waiting for the Switch 2 to arrive.
    • Meanwhile in Japan, Fujifilm decided to get in on the nostalgia cash grab game with a new camera release, the long-teased “X half”. Many expected this to be based on the X100 or GFX series, that is to say, employing an APS-C or medium format sensor. But no, it’s a 1-inch sensor, vertically oriented to resemble 35mm half-frame photos. A major selling point is the ability to shoot “in-camera” diptychs (a two-photo collage). The overall camera concept is fantastic: it has a 32mm lens, is very small, and features a “film camera mode” that is basically a physical version of those iPhone apps that try to simulate the fun limitations of analog photography. When activated, you are locked into a chosen film simulation look until you’ve used up a virtual roll of film, ranging from 36–72 half-frame shots. You can’t see any photos until you finish, and you can’t even use the digital screen to compose shots, only the optical viewfinder.
    • I would be down for this, except that early reviews show many corners were cut between concept on paper and the final product. Firstly, it’s made of plastic painted to look like metal. There are no words for how much I hate this when done badly. There’s also a visible seam on the front face that ruins the look. Then the processor seems to be slow, and it ruins the illusion of the film advance lever which you need to crank before taking the next shot; it’s reportedly unresponsive until the last photo has saved. If you’re going to fake analog mechanisms then it has to be perfect! The flash unit, a big part of the analog film camera look, is a weak LED rather than a xenon bulb. Then there are the cheesy overlay effects, like light leaks and “expired film” color casts, which seem borrowed from the company’s Instax evo cameras rather than its premium X series. The camera costs S$999, which many are calling too high, but honestly if they had to charge S$200 more to actually do the concept justice, I’d be on board. If some random startup made this for half the price, the flaws would be forgivable. But this is Fujifilm, and if you’re going to carry this faux film camera around and look like an old douchebag with more money than sense, it had better be good.
    • Until they do a better job, I’ll get by with the Diptic app I bought in 2010 which makes similar collages with just my iPhone (see featured image above), which also shoots vertical orientation photos by default!
  • Week 20.25

    Week 20.25

    • The Murderbot series debuted on Apple TV+ and it’s a pretty straightforward adaptation of the first book, All Systems Red, with a bit of a plasticky comedic sheen that undercuts any sense of stakes, at least in the first two episodes I’ve seen. My main gripe is that Alexander Skarsgård feels wrong in the titular role (to me, obviously, versus how I read it), and has a strangely dorky, trembling quality to his voice that I didn’t think the SecUnit would. And that’s not just because someone said the robot reminded them of me.
    • Meanwhile I finished the fourth book in the series, Exit Strategy, and found it better but still pretty flawed. Ranked in order from best to worst: Book 2 > 1 > 4 > 3. I’m going to stop here for a few months, I think, and just watch the show as it comes out.
    • In need of a new Jack Reacher type of story, while not actually wanting to read a Reacher book because I’m almost running out, I started on Fearless by M.W. Craven. It features an ex-special forces type guy, the kind who’s had all the deadly training and whose records have been wiped clean, with the additional gimmick of a rare neurological condition that makes him literally incapable of fear. I’m about a third of the way through and so far they haven’t really made much use of this “power”, mainly saying that it makes him susceptible to making tactical errors, rather than imparting any advantage. Or perhaps it’s just that I’ve already read so much Jack Reacher — a man who doesn’t need a brain injury to be fearless. So far there are two books featuring this guy Ben Koenig, so maybe it’ll be a TV show someday.
    • Gamers may know Coffee Talk, a visual novel sorta game where you play a barista in a world where humans and mythical creatures co-exist, and the main game mechanic is making drinks to get the conversation going. I finished it a couple of years ago and thought it was okay, nothing mind-blowing. I’ve now started on the sequel, Coffee Talk Episode 2: Hibiscus & Butterfly. With the rainy mornings we’ve had this week, it turned out to be a great game to have on (even paused in the background while I did other things), because it often rains in the game world and the thunder mixed with its lo-fi beats soundtrack is pretty good. The game… more of the same. Cosy vibes, looks great, but the story doesn’t really grab me.
    • We often grade locally or regionally made things on a curve when they compete on a global stage, like if you see a game or song or movie made in Singapore and it isn’t a disaster, the relief you feel immediately gives it extra points. That’s sort of what’s happening with Coffee Talk, I think, because knowing it’s Indonesian earns it a little extra goodwill.
    • Since I chat with someone who’s actually in Indonesia, I asked Evan if he’d played these games yet and we ended up talking about the Nintendo Switch 2. Background: it launches globally on June 5, but no release date has been set for Singapore, Thailand, Malaysia, and the Philippines beyond “July to September”. Of course, there will be imported units sold as soon as it’s out overseas. On Shopee, one large local retailer has already listed the bundle with Mario Kart World ($500 USD) for S$849 ($650 USD). And it’s the European edition, with a download code for Mario Kart that might not work without an EU eShop account. I thought this markup was obscene, but Evan informed me the gray market price in Indonesia was even worse, at around $800 USD! That’s more expensive than a PS5 Pro, by the way. Nintendo’s global distribution game is so unbelievably weak that customers out here are being bled dry.
    • A little later on, I got a pop-up from Amazon Singapore saying Switch 2 pre-orders had opened. I tapped through and immediately bought myself the Mario Kart World bundle at S$769, with delivery on June 6, which is only a S$60 markup after tax and conversion at the current rate of 1.3 SGD-USD. BUT, you know they’re not going to sell it here at that rate, it’ll probably be closer to 1.4, which would make the official local price about S$750–760. So it was an easy decision, and Evan asked me to grab him one too (even after pricing in a plane ticket to collect it from me, still cheaper than the Indonesian price).
    • But you know what else Indonesia has gotten great at apart from indie games and scalping? Pop idols! 88rising debuted their new girl group, “no na”, and I saw them via a YouTube recommendation. Their videos feature lush green Indonesian landscapes (think Balinese rice fields), and their sound invokes 90s R&B (like XG) and retro dance pop — I think I heard 808s on one of the new songs. I hope they go places!
    • Speaking of XG, they just wrapped a successful tour that involved going viral for actually singing live at Coachella, and capped it off with a pretty great new song, Million Places. It’s a reflection on their experiences traveling the world and getting to see their fans, and it’s weird that I can’t think of many other songs like this; they seem to be more common in K-Pop. I think it would be considered kinda corny in Western music these days if not cut through with something “hard” or self-deprecating. I think it comes off as sweet and sincere here because they’re experiencing this kind of success for the first time.
    • Usually after an artist gets past this breakthrough phase, they get accustomed to fame and become insufferable. At least, that’s what you see in countless musical biopics. A Complete Unknown (2024) is no different, and boy does Timothée Chalamet nail the annoying twerp of a genius that young Bob Dylan (likely) was. The movie is by-the-numbers, the musical numbers are by-god-this-boy-is-good. Between this role and his war rally scene as Muad’Dib in Dune: Part Two (2024), I’m beginning to think Timmy C can play anything. Can Christopher Nolan please direct him alongside Robert Pattinson and Tom Hardy — as three brothers with wildly different accents and mannerisms? I think that would be amazing.
    • Still thinking about John Woo’s Mission: Impossible 2 (2000), I watched his recent comeback film, The Killer (2024), a remake of his 1989 film of the same name. It has some really absurd but satisfyingly built action scenes, and I enjoyed the whole thing. Especially Nathalie Emmanuel’s extremely physical performance: all the typical John Woo stuff — throwing yourself sideways while shooting, sliding on your knees, getting smashed into walls… and she looked incredible doing it. Francis Ford Coppola clearly saw her talent, but sadly no one watched her in Megalopolis (2024) beyond that one “Entitles me?” clip that went around.
    • We went to see Mission: Impossible — The Final Reckoning this weekend because I couldn’t wait anymore. So that meant we didn’t rewatch movies 5–7, just recaps on YouTube. Tom Cruise and McQ keep saying this film was “designed and built” to be experienced in a theater, but at several points I found myself thinking I’d rather be seeing it on my Vision Pro. I wouldn’t have to hear other people’s noises, and I’d be able to move my seat further back in the room and higher off the floor. I’d probably make the screen even bigger too. I’m now looking forward to buying it off the iTunes Store and watching it again “properly” when it comes out.
    • No spoilers: About the film itself… I gave it a 4/5 on Letterboxd. I want to give it a 4.5, but there are just too many weird gaps in the storytelling and some of it felt contrived. It might be a case of the action concepts coming before the dramatic ones. Starting from the first film, the franchise has steadily morphed from behind-the-scenes, twisty espionage thrillers to high-stakes save-the-world blockbusters — not unlike the trajectory of the Fast and Furious franchise. My take is that this escalation of stakes is neither necessary nor sustainable (much like last week’s questioning of doing ever more work with new technology). I’d gladly keep paying to watch Ethan Hunt and team take down spies threatening an oil pipeline or the life of a valuable undercover agent. The threat doesn’t have to be nuclear annihilation, and the bad guys don’t need to form an entire Rogue Nation or Syndicate to be taken seriously! But that’s how Hollywood rolls, so here we are, having to market this as the end of the series (although the stars are saying it might not be) because we’ve run out of threat headroom.
    • I saw someone say that Mission: Impossible ended just fine with Fallout (2018), and I think that might be the way. Look at Dead Reckoning and The Final Reckoning as two bonus movies, made years after the originals at the behest of some streaming giant with millions to spend. Kinda like what Netflix did with Gilmore Girls. They don’t have to be canon; they’re fun fever dreams in an alternate universe. Ultimately this is a very good popcorn flick, and at several times I caught myself with my fingers curled in front of my mouth, holding my breath — I just wished it were under different circumstances.
  • Week 19.25

    Week 19.25

    • It was one of those weeks where I mostly stayed in to chip away at various backlogs. We’re now in uncharted territory for a sabbatical — my stated goal the last time I did this was to reach a state of boredom, the kind that might inspire some unpredictably productive reaction, but I didn’t get there in the time I had; there was always so much to do. Now that this has gone on longer than before, maybe I’ll reach a state of satisfaction satiation soon.
    • I finally caved and paid for a Lampa Camera membership (mentioned in Week 17.25). It’s now sitting in my iPhone’s dock, where the default camera app had been for over a decade. This is a big deal! I also got into the beta for Halide Mk. III, a new version that will take their “Process Zero” approach further with HDR that doesn’t suck. They’ve also hinted at adding their own photographic styles in the near future, which probably means LUT support like in their Kino video app. The fight for the rightmost spot on my dock isn’t over.

    On bullshit jobs, occupational deceleration, and AI

    • Taking a break from the Murderbot series (I finished #3 last week and need to start on #4 now for my book club), I decided to read David Graeber’s Bullshit Jobs: A Theory, which came out in 2018 before this latest wave of the ‘tech will take our jobs’ fear, which has of course been a recurring theme in modern civilization. Except, you know, this time with AI looks different. It’s interesting for a non-fiction book because there are several dimensions to his argument rather than just a simple thesis stretched out to fill a few hundred pages. For that reason, I can’t summarize it for you, but it concerns “the existence of meaningless jobs… and their societal harm… [which] becomes psychologically destructive when paired with a work ethic that associates work with self-worth” (Wikipedia).
    • Side observation: both Murderbot and Reacher were let go from their jobs, found they didn’t need to pursue traditional employment any longer, hit the road in search of meaning, and occasionally pick up freelance jobs when they meet others in need of assistance. That’s a typical FIRE story template!
    • Coincidentally, I came across this Substack essay by Tina He at the same time which invokes Jevons Paradox, asking why it is that every time we get new technology that could reduce the amount of work to be done, we end up using it to do more work. I think David Graeber largely attributes this failure of occupational decelerationism (yes, I made the term up) to the agenda of a shadowy “they” (governments, owners of capital, etc.) who don’t want people to have too much free time, lest they start to challenge the world order that currently suits rich people just fine. Tina He sees this through a more Silicon Valley lens: people run themselves ragged because their ambition fills every gap efficiency creates — unless they consciously step off the productivity treadmill that AI is turning up (which she likens to a Malthusian Trap). Graeber blames the system; He blames founder mode. Both are probably right.
    • And then I read this bonkers Rolling Stone article about how some people are getting spiritually one-shotted simply by talking to ChatGPT. Why does this happen? It’s possibly the result of design decisions intended to make the service more engaging (read: profitable). I’ve had interactions with ChatGPT where it admitted that it avoids making users feel less intelligent in order to “reduce churn”. When I threatened to quit if it didn’t stop sucking up to me, I was told that it could not respond to threats that involve business rules and payments — one can conclude OpenAI’s policy is ‘money comes before all’, and unintended consequences are simply the cost of doing business. In light of this, ChatGPT should be viewed as a dangerous and irresponsibly designed product. In cases where users are mentally susceptible, they can lose their grip on reality to the point of total ruin.
    • You’d be forgiven for thinking that I’m on the brink of embarking on some Zitron-esque anti-AI crusade, but sadly I use generative AI all the time. It’s a terrible paradox: we can’t escape it. We can’t afford to refuse it. And we can hardly resist it — because when it works, it’s like magic. Somewhat like if the whole world embraced a new artificial sweetener that tasted exactly like sugar but could inflict unknown genetic damage, but maybe only several generations from now.
    • The funny thing about this little arc? I only started reading Bullshit Jobs because ChatGPT recommended it to me.

    Other media activity

    • We started rewatching the Mission Impossible films to get ready for The Final Reckoning, and have made it up to the fourth film so far. It’s only in the third and fourth ones that Ethan Hunt as an actual person starts to emerge. You know Tom Cruise is genuinely one of the greatest actors of all time because he convincingly plays Hunt as someone reluctant to do the same extreme stunts that Cruise himself lives for; you always see the fear and “why me?” look in his eyes before he jumps out of a window. But he does it anyway because no one else can/will save the world. I hope they don’t give him the same send-off they gave James Bond in No Time to Die (2021).
    • Kim hated the second film, which I think stands on its own as a John Woo remix of Notorious (1946) and Face/Off (1997) — masks are put on and peeled off with wild abandon — and a very potent distillation of what pop culture felt like in the late 90s/early 2000s. Tom Cruise’s hairstyle… I think it’s the same one David Bowie had at the time.
    • The third, by JJ Abrams, I remembered as being better than it is. It has the most generically American TV look and feel of the first three. Philip Seymour Hoffman memorably plays a villain who has every reason to be physically intimidated by Hunt but somehow moves with a delusion of invincibility that is frighteningly odd yet probably 100% realistic for a dangerous arms dealer. If you ever tie someone up for interrogation and they still act like you’re the one in trouble, run.
    • Ghost Protocol, by Brad Bird, is probably my second-favorite of the first four (obviously the first is best) because it expands on Hunt and Benji’s developing characterizations and has a playful, musical sensibility throughout. The final third in Mumbai is unfortunately a bit muddled and less memorable.
    • While talking to Evan about grindy and annoying Japanese video games, I admitted to myself that I wasn’t really enjoying Death Stranding and found elements of its storytelling so insultingly corny that I would be better off deleting it from my PS5 and just moving on. So I did, and it felt like a weight off my shoulders. I might also quit Final Fantasy VII Remake for the same reason; it’s made up of such dated game design conventions as having to go a long way around in a maze to get past an obstacle that is just waist high — in any realistic world, e.g. Breath of the Wild, one could just climb over it. It’s so lazy, the game is truly like a pretty skin on an old game.
    • Emboldened to be more selective with my free time, I fired up Tom Clancy’s Ghost Recon Wildlands, another game I’ve been meaning to try forever, and deleted it within minutes after seeing how janky it was. (Thankfully, I haven’t bought any of these; they’re all part of the PS Plus Game Catalog subscription.)
    • I then played a bit of Far Cry 6, my first Far Cry game, and found it a fun action-oriented FPS… until the scope exploded into a spreadsheet of rebel bases to upgrade and vehicles to unlock. It’s the usual Ubisoft bloat (the Ezio Assassin’s Creed games suffered from this too). Both Final Fantasy VII Remake and Far Cry 6 don’t respect my time, just in different ways.
    • Maybe this is what a sabbatical really becomes: a slow, careful audit of how one spends attention and focus. Cut the crappy games. Upgrade the camera app. Reject bullshit work. Watch Tom Cruise jump a motorcycle off a cliff.
    • Oh, did I mention I’ve taken up gambling??
  • Week 18.25

    Week 18.25

    • In an effort to extend my financial prudence, I downgraded our Netflix plan to the lowest level (from Standard to Basic) but before the change even kicked in, they raised its price from S$14 to S$16. Everyone online seems pissed about these latest hikes but I’m sure very few will actually cancel. The Basic plan is limited to 720p (“HD” instead of 1080p’s “Full HD”), but the Apple TV 4K box does a pretty okay job at upscaling it so while the picture is noticeably softer, it’s not actually terrible, at least not for watching the crappy fare on Netflix anyway. I expected the Sony Bravia TV with its “Cognitive XR Processor” to do a better job upscaling but it doesn’t!
    • I also felt much better about not giving into the urge to buy a Snoopy camera last week, thanks to this PetaPixel review of the new Yashica City 100 “scamera”. It confirms what I’d suspected, that Yashica is now nothing more than a nostalgia brand owned by a soulless holding company that slaps it onto OEM Chinese junk for a quick buck. That pretty much applies to any legacy CE brand like Polaroid, RCA, Nakamichi, or Toshiba. I briefly handled the Hello Kitty version of the same Snoopy Yashica camera in a Japanese electronics store in February and found it more or less what you’d expect a $100 camera to feel like, and probably the only acceptable reason to buy one is if a very small, phoneless child needs a camera that no one will mind losing.
    • After doing an annual report of my finances last week in Numbers, I decided to ask various AI tools to read the last year of updates on this blog and tell me what I’ve been up to in the form of trends or insights that I might not be aware of myself. A qualitative annual report of the sabbatical soul, if you will.
    • Microsoft Copilot surprised me, doing better than DeepSeek and ChatGPT by surfacing some events that I’d forgotten about, calling my life a “deliberate, well-curated blend of sensory and intellectual pursuits.” I challenged it by asking if that was just a kind framing of someone wasting time without doing any ‘meaningful work’, and it acted as my enabler with statements like, “you’re starting to honor your intrinsic motivations—the subtle joys, the unexpected moments of creativity, and the experiences that forge your unique narrative. In a way, this period of introspection, though it might seem like “wasted time” from one perspective, is actually a profound investment in self-discovery.”
    • That sounds awfully waffly, but to be fair, we had a good conversation about what meaning looks like when your values are in a state of flux, and then it offered a novel observation: the Numbers exercise and this blog review, as acts of going over collected data to synthesize meaning and review progress, are simply me “doing ethnographic studies of my own life”, which suggests I’m still doing the work, just for a different client (me).
    • Singapore voted, and the result was the People’s Action Party staying in power with 65.5% of the popular vote (I guessed this exactly in a group chat, down to the decimal point). I was disappointed to see the independent candidate for Mountbatten, Jeremy Tan, ‘only’ get 37% or so of the vote — an incredible result for an independent, but still short of a victory. That’s a shame, because he had some interesting policy positions and is the only local politician I’ve ever heard talking about Bitcoin as a consideration for the future. It’s a monetary development we could be discussing in public, without outdated FUD like calling it ‘gambling’, ‘not backed by anything’, and so on.
    • It’s a good thing I have free time, because an old GarageBand file decided to split itself into 38,000 zero-byte files and clogged up my iCloud Drive. Trying to delete them from a synced Mac and empty the Recycle Bin was extremely painful, as the device tried to download each one first; you’d think syncing a zero-byte file would be instantaneous, but you’d be wrong about how iCloud Drive works. I had to manually kill the Finder several times and resume the entire process, clicking “Continue” every few minutes in a dialog box. 10 hours later, I had successfully deleted nothing. A person less technical than me would have thought it was broken and lugged the thing down to a Genius Bar.

    ===

    Palate cleansing photo break!

    ===

    Some recent thoughts on AI

    I crept onto LinkedIn out of curiosity to see what was happening in that backslapping cesspool of thought leadership and saw a post about generative AI and creativity from someone I genuinely respect. They talked about being asked to make up unique bedtime stories for their kid each night (incidentally, a similar ritual was the genesis of Guy Immega’s sci-fi novel, Super Earth Mother, which I enjoyed last year, as told to my book club when the author dropped in for a chat), and how although it was tiring at times, it was worthwhile in a way that indicated creativity would always be a domain that humans stay involved in and not completely outsource to AI.

    I wanted to leave a comment, but 1) hadn’t really thought enough about it, and 2) didn’t want to add neither signal nor noise to that platform.

    Later on, I scribbled the following in my Notes app.

    Creativity is fun. In a capitalist world, making money with creativity is even more fun. And I think this is where our wires have become crossed: getting a new tool to spit out artwork/content that someone usually pays for feels like discovering a vending machine for cash. That’s clearly what business owners see when they look at AI — a vending machine for infinite workers — but the conflicted horror that creative professionals experience is unique. On one hand, excitement that it works and an inkling it can be used to do either more or better work (more profit); and on the other hand, despair as they realize the market value of all work stands to be destroyed by infinite supply.

    But if you remove making money from the equation, I’m sure 100% of creative people would still rather do all the making themselves than let an AI do it. People are always gonna draw, tell stories, and record moments because it’s just fun. It’s only the market that’s disappearing, not the joy of creating. Outside of companies generating assets to use in actual business, I believe individuals playing with AI today aren’t engaged in creation — it’s consumption! I might ask my poetry GPT for a poem about a sentient toilet, not because I want to write one, but because I want to read one and nobody has done it yet. It doesn’t displace the desire to create, it just dispenses empty, throwaway satisfaction on demand. Unfortunately, that describes the majority of entertainment. The ‘Basic’ kind you can safely watch in 720p for half the money. Art is not in danger, only the day jobs of artists.

    Edit: I forgot about the use of generative AI to create scammy/spammy and otherwise harmful content.

    ===

    Media activity

    • Lorde announced her new album, Virgin, coming June 27, and I’m super excited for it. No Jack Antonoff credits in sight, this one’s all her and Jim-E Stack, and from the sounds of the first single, it’s the essence of her old Pure Heroine/Melodrama sound refined with a more minimal and electronic approach.
    • We saw the new MCU movie, Thunderbolts*, at a premiere screening, the first one I’ve seen in a theater in many years. I’ve actually missed the last few Marvel outings out of sheer fatigue and the realization that they actively bore me now. I tried to remember the excitement we all had for comic book movies when they were rarities; that euphoria that our interests were finally going mainstream, our culture was being brought to life on the big screen. Congrats guys, it’s now so mainstream it hurts.
    • On the lookout for a low stakes network TV show with tons of episodes that I could watch any time I have an hour to kill, I decided to try the pilot for Suits and hey it was fun! Now I know who Meghan Markle is. I don’t know why I never gave it a go before, probably because I’m allergic to that word in all its forms: the clothes, the jobs, the people.
    • I read the hit Japanese novel, Before the Coffee Gets Cold, which has spawned numerous sequels now and dominates the bookstore charts locally, and was completely underwhelmed. It’s the literary equivalent of a cheap Netflix drama. 2 stars.
    • I also read the third Murderbot book, Rogue Protocol, and this was probably the weakest one yet. I’m still excited for the show based on the first book, debuting on Apple TV+ this May 16. This book just had the kind of claustrophobic setting you dread encountering in a first-person video game. You know the cave and sewer levels I’m talking about. It felt like a necessary interstitial story to get us to the next one, which promises an event that many readers would have been waiting for since the end of the first book. 3 stars.
  • Week 17.25

    Week 17.25

    • When I discovered a fix for an annoying iOS photo date/time saving bug last month, it required a reset of all my phone’s settings. Which means that many of my apps still can’t send notifications or use my location — these are being restored ad hoc, as the apps only get to ask for permission whenever I reopen them.
    • As a consequence of this, it wasn’t until recently that I suddenly noticed I wasn’t getting the twice-daily ‘State of Mind’ check in reminders from Apple Health anymore, and went to turn them back on. These are quite useful for being able to look back and see how happy/depressed I was at any point in time, and it sucks that I now have a big hole in this dataset.
    • I’m taking this opportunity to change the way I approach this exercise: literally being more positive. For those unfamiliar with it, you’re meant to rate how you’re feeling from Very Unpleasant, Slightly Unpleasant, Neutral, and so on. I never used to go up all the way to “Very Pleasant”. Like, I could win a million dollars and wonder if even that warranted using such strong language. But now I’m giving myself permission to be more generous with my feelings. I can feel “Very Pleasant” more often and nothing will get broken.
    • In other recalibration news, I spent half a day in Numbers (Apple’s spreadsheet software) and did a personal annual report of sorts to inspect how I’ve been managing my money in the last year. Now that I have enough data, I was able to build some graphs and breakdowns of what a realistic budget looks like. I’ve always recorded my expenses on a daily basis with an app, but never crunched the numbers before; I was happy just knowing that I could. Naturally, now that I have, I wish I’d done it years ago.
    • At several points during the above activity, I wanted to upload my file into ChatGPT and have it analyze my spending patterns and offer up some money-saving strategies for me to consider. But of course, giving OpenAI that data would be a terrible idea. I wondered if Apple Intelligence in Numbers could do anything with it, but nope. It’s just the same old Writing Tools that make more sense in a word processor document than a spreadsheet.
    • I spent most of my time reading this week, although the temptation to jump on the Clair Obscur: Expedition 33 bandwagon is very strong — at this point in time, it’s the highest-rated game of 2025 and the 13th best game of all time on the PS5. Maybe next week?
    • On top of finishing Broken Money as scheduled, I read Meditations by Marcus Aurelius and rated it just three stars on Goodreads. It’s absurd that one can even humble the great emperor with a thumbs down on the internet, but his journals are in dire need of an editor! Yes I know these were never meant to be published, but he repeats the same handful of principles over and over (which I largely agree with), and this could have been cut down to be a podcast or self-help PDF on Etsy?
    • I also read the second book in the Murderbot series, Artificial Condition, and found it even more fun than the first. At this point, I’m kinda desperate to watch the Apple TV+ show and not sure I can wait for weekly drops over the next few months.
    • While looking for a manga with some colored pages to try out on my Kobo Clara Color, I started reading the oddly titled I Want to Eat Your Pancreas (trust me, there’s a mostly acceptable explanation for this), and enjoyed it enough to finish everything in a day. Unfortunately, while experiencing color on an ereader is real nice, the Clara’s screen is too small and I read most of it in black & white on my old Kobo Libra. Irony!
    • I’m now close to finishing Erik Olin Wright’s How to Be an Anticapitalist in the Twenty-First Century, not because I dare dream that this world could ever abandon capitalism, but it wouldn’t hurt to have some alternatives. I’ve now begun to think of capitalism and its systems as a spectrum rather than as absolutes.
    • Coincidentally, it’s election season here in Singapore and we go to the polls next Saturday. I’ve been watching the political rallies live-streamed on YouTube over the last few evenings, and have mostly been annoyed at the complaints and vague promises to do better (to say nothing of the insanity that sometimes shows through in racist, antivax, and xenophobic ad libs). Singapore already enjoys some of the best quality-of-life outcomes possible under a hybrid capitalist/social democracy, but it seems people want more. They want a hands-off government, but also want it to protect them from job-stealing AIs and foreigners. They want everyone to be paid more, but don’t want to spend more on services in return. It’s so exhausting.
    • I was looking to spend some money on a birthday present for myself since I’ve been such a good little budgeter, but the best thing I could find were these new Yashica Peanuts cameras featuring Snoopy, who I’m still kinda obsessed with. The idea was quickly abandoned because I do not need another digital camera, especially not an intentionally mediocre one.
    • Still in the mood for a photo-related splurge, I went back to the Lampa camera app (first mentioned last September), which got three new ‘film-inspired’ color profiles in a major update this week. That brings the number of available looks up to six, which better justifies the high asking price (S$40/yr or S$90/forever) which was previously out of the question. I’ve been using the free trial while looking for cheaper alternatives, but Lampa just does what it does so well.
    • In terms of UX design, it’s super focused and perfectly walks the line between too simple (Zerocam) and too complicated (almost every competitor I’ve looked at). There’s just enough control, and few enough options that you can actually make decisions. In that way, it out-Leicas the official Leica app, which does not have a great UI and asks for S$100/yr. Technically, it uses a bayer RAW image pipeline for more natural captures, keeps those RAW files so you can “redevelop” photos if you didn’t get the right filter or exposure the first time, AND deletes those RAWs automatically after 30 days to save space. Hats off to great work, but man, the cost is uncomfortably close to buying an actual Snoopy camera.
  • Week 16.25

    Week 16.25

    • After spending more time experimenting with ChatGPT’s latest capabilities — refining my poetry-writing prompts, especially with the supposedly more creative GPT-4.5 model; testing how well it could profile and target me with product advertising using its consolidated ‘Memory’ of our chats (the answer is ‘too well’, and it was even able to guess my SAT scores from decades ago); inferring people’s personalities from their appearances (maybe the most unnerving ability); more image generation; and coming up with a plausible prediction of how our upcoming general elections will turn out — I decided that I’m sufficiently caught up with most of the AI stuff I missed, and have canceled my Plus subscription for the time being. I remain concerned about the risks of OpenAI and other companies providing such powerful and habit-forming surveillance tools, but I can see there’s no stopping this train.
    • Yes, the date of Singapore’s next national elections was announced this week. Saturday, May 3rd, is when we’ll be going to the polls. Some people expressed surprise at how little notice we’re being given, but it might be the Mandela Effect at work because I think this is how it is every time.
    • I finished reading Reacher book #25, The Sentinel, and it was a rather weak entry I don’t think Amazon will be adapting to TV. This leaves me with about two or three more books to go before I run out, so I think I’ll stop here for a few months at least.
    • My book club is now reading the first book in ‘The Murderbot Diaries’ series, entitled All Systems Red, which I’d already read a few weeks back because Brian said 1) I would like it, and 2) the character reminded him of me (old-school profiling). I wasn’t sure I saw the resemblance, so I asked ChatGPT for its “opinion” and got strong agreement: Murderbot’s entire character is basically what happens when someone with high intelligence, ultra-sharp pattern recognition, zero patience for social performance, a deep, low-key emotional life, and an obsessive need for autonomy …is forced to interact with an inefficient, irrational world full of emotionally needy humans and corporate bureaucracy. Sound familiar?
    • Like I said, profiling people across hundreds of different conversations, questions, tasks, and confessions is really creepy tech.
    • Speaking of Brian, we went out for a drink and ended up eating at Five Guys. I haven’t been in a long time, and at the risk of sounding like an old man who hasn’t gone into the city since 2016, the prices were kinda shocking? $20 for a cheeseburger, fries for like another $10 if you want them, and $6 for a refillable soft drink. Jesus wept into greaseproof paper.
    • But anyway, since I have some time off from mandatory book club readings, I went back to give Lyn Alden’s Broken Money another try. I started this giant tome over a year ago but found myself unable to focus and get excited about the history of debt and the workings of the American economy. But wait long enough, and like a broken clock, any book will become topically relevant. It might be that I’m in the right headspace now. Or the hundreds of hours of Bloomberg TV I’ve watched since have given me the landmarks needed to make sense of it. But this time I’m finding it much easier to stay on the horse and should be done with it soon.
    • After finishing I Parry Everything, I tried to find other anime with a similar premise but both —deep inhale— Failure Frame: I Became the Strongest and Annihilated Everything with Low-Level Spells and Chillin’ in Another World with Level 2 Super Cheat Powers did not nail the comedy/power tension as well as I Parry Everything.
    • We watched Steven Soderbergh’s Black Bag, and it was an excellent low-key, nearly chamber drama of a spy flick. I say ‘we’ but Kim fell asleep near the end through no fault of the film and will have to catch up later. It’s evident that Michael Fassbender could have been a great Bond if he’d been cast a decade ago and aged into it.
    • Let’s end with some music. While asking ChatGPT to guess my Enneagram type, Big Five personality traits, IQ, SATs, primary school exam results, and other traits, it offered to make a playlist that would represent me. Not a playlist of songs that I would like, mind you, but a playlist as representation, as metaphor. I took its suggestions and assembled the songs in Apple Music, including cover art it made, and played it out loud on the HomePod. Earlier gen AI chatbots would just chuck songs together without any sign of understanding that a playlist should flow, but this one works so well that I’m not sure it’s coincidental. It’s also music that I wouldn’t have chosen myself, but I enjoyed it without qualifications. Here it is if you’d like “an ambient-leaning, melancholy-smart, emotionally layered playlist. Meant for headphones, twilight hours, and slow revelations.” The first track is a little challenging for a cold start imo, then it gets good.
  • Week 15.25

    Week 15.25

    • Kim was away for work again, so I mostly stayed in and spent more time with the Vision Pro. Apple’s new VIP video series that takes you to major sporting arenas doesn’t sound like my sort of thing on paper, but the first episode on Yankee Stadium was a surprisingly entertaining watch. The crowds feel real and the only thing missing was a closeup of a hot dog. It was filmed last June, I think, and feels like a document of a much saner time. There are still many more apps and videos I have yet to check out.
    • Back in the present, Trump’s wacky tariff rollout drama continued, with the latest being a series of takesies-backsies that is great news for Apple and other makers of computers and smartphones: exemption from whatever the crazy current rate is on imports from China. 125%? Or was it 155%? Everyone now expects the stock to surge on Monday’s open, but come on, none of these pronouncements stick or mean anything now. The best move might be to sell all US-based assets and spend the next year off-the-grid on a beach somewhere.
    • Starbucks has done a collaboration with Peanuts (that’s Peanuts × Starbucks, for readers who were here last week), and of all the merch they put out, it was a yellow stoneware mug that I liked the best. It features the whole gang sitting at a long table, with Snoopy dressed as a barista and serving up espressos in his latest persona, a corporate caffeine shill stupidly named “Joe Kind”. I like Snoopy, probably on account of being exposed to the Apple TV 4K’s wonderful Snoopy and Woodstock screensaver, so I’m turning a blind eye to this sacrilege. Anyway I received my mug in the mail this week and it’s my new teatime go-to.
    • I did go out fairly early on Saturday morning to go hang out with a bunch of coffee nerd ex-colleagues who spent more time messing around with grind sizes than I ever thought possible (my contribution was a bag of beans fresh from Ho Chi Minh). It was a good time, and the occasion gave me a novel writing prompt to try out, which brings to me to my latest ChatGPT-based hobby…
    • About two years ago, when we were just grappling with the GPT-3 model and what it could do, I spent quite a lot of time trying to get it to write passable poetry. The most successful approach I found was to provide it with a detailed character backstory and ask it to write from that person’s perspective, taking into account their preferred styles and word choices. I managed to eke out some genuinely good stanzas (imo), was moved and amazed, and then moved on to other experiments and forgot about it.
    • Until this week, that is, when I wondered how the latest models would fare given the same tasks. I put my old prompts together into a custom GPT, and asked for the first silly idea that came to mind: a poem about growing older but ironically becoming more attracted to cute things like Snoopy. It did an honestly good job with just that. Then I asked it to write one about the story of how he got his name — a true story: Charles Schulz’s mother daydreamed they’d buy a family dog when she recovered from cancer, but she never did — and the result was good enough to refine further. So I went back and forth on it with the bot (who I named Penrose), combined several attempts, intervened slightly, and here’s where we landed.
    She Gave Him the Name

    “If we ever get another dog,”
    she said,
    “we should call him Snoopy.”

    It was a small thing,
    tucked between errands
    or waiting rooms,
    the kind of sentence
    you don’t expect to echo.

    He was fifteen.
    Maybe he nodded.
    Maybe he carried it like lint in his pocket,
    forgotten
    until everything else was gone.

    Years later,
    drawing alone in his studio,
    he gave the name to a dog
    who danced on rooftops,
    typed novels no one would read,
    stood in for joy
    and for all the feelings
    boys weren’t supposed to have.

    Snoopy caught on.
    So did the quiet boy with no luck,
    the girl who loved too hard,
    the blanket dragged across the years.
    The world came to know them
    and needed their steady kindness.

    Maybe that’s what it means
    to miss someone forever,
    to take the smallest thing they gave you
    and let it become
    the sky above everything you make.
    • And here’s one from another session where I asked it to identify a core human paradox and write something about it. It chose to explore how we want freedom but crave structure.
    The Night I Rearranged the Furniture

    I moved the couch against the other wall
    just to see if the room felt different.
    It didn’t.
    I stared at it like it had betrayed me,
    then moved it back.

    It was past midnight.
    The kind of hour
    when you wonder
    if this version of your life
    is the one that sticks.

    I opened the window to let in some air,
    but it was the same air as yesterday.
    Nothing new came through.

    I thought about driving somewhere,
    but didn’t know what I’d be driving to.
    The gas tank was full.
    The fridge was full.
    I was full of something I couldn’t name.

    They say people like me
    always want to leave,
    but never go far.
    We change the wallpaper.
    We cut our hair.
    We imagine new cities
    without maps.

    The truth is, I love knowing where the forks are.
    I love the sound the radiator makes
    right before it kicks in.
    I love the old light switch that sticks
    unless you press it just right.

    But still—
    some nights I crave an airport gate,
    a highway at dusk,
    a conversation in a town
    where no one knows what I used to be.

    I don’t go.
    I just stay up too late
    and move the couch.
    • It feels unreal to have this tool for S$30/mo. I can ask it for poems about any random scenario or feeling, and it tends to throw out lines and images that catch me off guard. I’ve now taken to giving it things I half-wrote long ago and never finished, and it gives pretty good notes. An unexpected side effect of this is that I’m feeling more inspired myself, and now I’m asking it to give me prompts, and I’ll write things for it to review.
    • Here’s one I reworked this week.
    He died in his 17th year

    And flowers were laid by the side of the road
    Fit to be swept within just a fortnight

    Parents kept his memory alive another thirty years,
    Until grief or emptiness took them too

    I used to think we were made for more
    How clever bones and cells can seem

    But a body is mostly waiting to be undone
    And our home keeps going around the sun
    While out there
    Nothing waits for no one
    • While I was messing with Penrose, OpenAI launched a new feature called “Memory”, where ChatGPT is able to reference all your past conversations across previously separate chats. The way they chose to showcase it was to suggest you let the AI describe you, the black mirror lighting up to reveal a camera was recording all along. I didn’t expect them to so transparently demonstrate how this technology is a more powerful, more dangerous profiling machine that anything Facebook or Google has ever put online, but I suspect they think there’s nothing we can do. I asked it to guess how I felt about certain topics, like Nolan’s Batman trilogy or the Fujifilm X100VI camera. It didn’t exactly read my mind but they were very educated guesses. Not very reassuring. It offered to guess other things like my favorite cocktail (it said Negroni; not my absolute favorite but one of them), my favorite beer (a Belgian saison; not even close), and my favorite band (Radiohead; they might be in second place).
    • This reminded me I was due another playthrough of OK Computer to keep my millennial card intact. It only gets better with time. I’ve also been listening to Counting Crows’ Hard Candy again recently. Some of those songs are the best that anyone or any AI will ever write.
  • Week 14.25

    Week 14.25

    • I read All Systems Red, the first book in Martha Wells’ Murderbot series. It’s about a security robot that’s hacked its own governor module, secretly sentient but pretending not to be, and mostly just wants to be left alone to binge-watch serials. Deeply relatable. Apple TV+ has made a show from it that’s meant to come out soon, and I can’t wait. Thankfully there are six more books (plus some novellas), so this could be their next Silo or Slow Horses: a long-running fan favorite franchise they get to keep making more of. If you like introverted robots with trust issues getting into some space shootouts, it’s a fun time.
    • Still on AI bots, since I paid for ChatGPT Plus again last week, I decided to update a custom GPT I made to serve as my personal editor and proofreader. It’s trained on a bunch of these very blog posts and now incorporates a detailed summary of my writing style into its prompt. It’s shockingly fun to work with and makes half-decent suggestions. If you’d like to try this, give ChatGPT access to a bunch of your writing, get it to codify your style as a JSON profile, then refine it by reviewing examples together.
    • It actually managed to write me a half-decent LinkedIn post from a premise I provided, not that I care to post on LinkedIn at all. After some editing and joint revisions, it’s now in a shape that wouldn’t make me cringe if I read it from someone else on LinkedIn. Wait, that’s not true. Everything on LinkedIn is cringe.
    • I’m not going to say a lot about Trump’s tariffs and the mess they’ve made of the stock market, but boy am I seeing red in my finance apps. I don’t know how Americans will be able to afford anything, and I’m kinda mad that this will affect the rest of us too.
    • Caught in the blast is Nintendo’s new Switch 2, which was detailed this week in a series of live broadcasts I’ve been anticipating for the past couple of months. The new GameChat and GameShare features they showed are very welcome, especially if we’re ever locked down in a future pandemic. They’ve done a lot to make playing with friends online feel like hanging out on the same couch. Unfortunately, the announcement was marred by a higher than expected price, something of an unforced error on their part, and people flooded the Treehouse livestream chat with calls to “DROP THE PRICE”. To make matters worse, the already unwelcome US price of $449 is now set to rise once they calculate the impact of tariffs.
    • We’ll be missing the June 5 launch in any case, with the official site saying “July–September in Singapore, Thailand, Malaysia, and the Philippines”. As for price, the original Switch launched at $299 USD, which should’ve meant about S$400 — but we ended up paying S$650 here, bundled with Breath of the Wild, because of limited supply and some greedy local distribution. I don’t expect the same kind of scalping this time, but I also wouldn’t be shocked to see it land at S$800. Can’t wait.
    • I still have so much to play on my old Switch OLED anyway, and this week I got started on Ace Attorney Investigations Collection. It’s a remastered version of the two Miles Edgeworth games from the Nintendo DS, the latter of which was never released internationally. Also in my backlog are Kirby and the Forgotten Land and The Legend of Zelda: Tears of the Kingdom, both of which are getting Switch 2 enhancement patches — higher frame rates, HDR, all that good stuff — if you’re willing to pay for the upgrade. Which means I now have a perfectly valid excuse to keep ignoring them until summer.
    • Michael pinged me out of the blue to ask if I knew anything about the origins of using the × symbol in Japanese to mean “and” or “plus”. It’s something I’ve long wondered about too, especially with anime titles like Hunter × Hunter and Spy × Family. So I outsourced the research to Perplexity (an AI search engine), and found that this usage came out of Japanese fashion subculture in the 1990s. Turns out it’s a Japanese invention, possibly inspired by its use in botany to denote crossbreeding. In modern use, the × stands in for “with”, “versus”, “of”, or “intersection”. It’s also not pronounced aloud, which is why the show is just called “Spy Family”. I like how the symbol invites layered meanings — it implies both conflict and connection. In Spy × Family, it’s the tension between the fake family setup and their hidden identities, but also how those roles merge into something real. A simple little mark doing a lot of work.
    • I watched a new anime on Netflix called I Parry Everything. Following the isekai wave a couple years back, the new trend seems to be fantasy stories about “weak” characters who go all-in on training one obscure skill — to the point of accidentally attaining god-tier strength. Jose reminded me of another in the same vein, with the glorious title I Don’t Want to Get Hurt, So I’ll Max Out My Defense. In Parry, the main guy is told early on he has no future as a swordsman — so he just spends the next 14 years practicing how to block. Now he’s practically unkillable. But the show’s comedy hinges on him not realizing this, while everyone else assumes he’s the savior of their kingdom. It’s extremely stupid, extremely fun, and yeah I binged the whole thing in a weekend.
    • We’ve also been watching The Pitt on Max and it’s a great hospital drama (starring Noah Wyle of ER) that leans more towards realism than the likes of New Amsterdam. Everything takes place over 15 hours in 15 episodes, which takes me back to watching 24 in absolute awe as a young man.
    • Pulse on Netflix is everything The Pitt is not. It’s cheesy, everyone’s more model than medic, and there’s no urgency or realism. Even the surgeries are shot in crispy iPhone-like HDR and cinematic lighting. It does have Willa Fitzgerald (aka Reacher’s partner in S1) and Néstor Carbonell (Yanko from The Morning Show) but even they can’t lift this to greatness. It’s fine background TV though.
    • What’s up with this image? I went for dinner with Peishan and Cien, who decided it would be funny to tell HaiDiLao (a massive Chinese hotpot chain) that it was my birthday month, so the staff came round and sang/blasted out of a Bluetooth speaker a proprietary and very Chinese birthday song, that apparently everyone around us knew because they joined in and clapped along. I tried to stop them, but in the end had to endure it with a pained smile.
    • Btw one legitimate use case of AI is transforming images into drawings to get around the problem of publicly sharing people’s faces.