Someone Trained a Tiny AI to Pretend It Is a Terminal...
Simon Edwardsson built a terminal with nothing behind it. His WorldTUI is a 360-million-parameter language model that takes each keypress and predicts how the screen should change, with no shell and no filesystem n his blog (October 21). It still runs vim and tmux, plays snake, and usually remembers what you wrote a few commands earlier.
The trick itself is older than you'd think, so the size is the impressive part. In December 2022, Jonas Degrave wrote about prompting ChatGPT into behaving like a whole virtual machine. Google researchers later did the same kind of thing for Doom with GameNGen, which simulates the game at over 20 frames per second on a single TPU. Edwardsson trained his version in about 75 GPU-hours on one RTX 4090, and it ships as a 208 MB download that runs in a browser tab.
Memory is where it pulls ahead of the general models he compared it to. On his file test, WorldTUI read back 9 of 13 files right, against 2 and 3 for two prompted Qwen models. He calls his snake "probably the most compute-intensive snake in the world," and I can't argue with him.
He's honest about where it falls apart. Anything harder than simple Python, he says, produces answers that remind him of GPT-3-era hallucinations, and nested directories can drop you into a different path. Hacker News gave the post 463 points, and one reader asked the question I'd have asked: does it recognize a row of asterisks as a password?
If you want to play with it, I'd use Chrome or Edge on a desktop with a real keyboard, which is where Edwardsson says it works best. Type pythoon with the typo and it launches Python anyway, which is how you know nothing real is listening.
Is it useful? His own answer is "Not really." That's why I'd try it.
Simon Edwardsson built a terminal with nothing behind it. His WorldTUI is a 360-million-parameter language model that takes each keypress and predicts how the screen should change, with no shell and no filesystem n his blog (October 21). It still runs vim and tmux, plays snake, and usually remembers what you wrote a few commands earlier.
The trick itself is older than you'd think, so the size is the impressive part. In December 2022, Jonas Degrave wrote about prompting ChatGPT into behaving like a whole virtual machine. Google researchers later did the same kind of thing for Doom with GameNGen, which simulates the game at over 20 frames per second on a single TPU. Edwardsson trained his version in about 75 GPU-hours on one RTX 4090, and it ships as a 208 MB download that runs in a browser tab.
Memory is where it pulls ahead of the general models he compared it to. On his file test, WorldTUI read back 9 of 13 files right, against 2 and 3 for two prompted Qwen models. He calls his snake "probably the most compute-intensive snake in the world," and I can't argue with him.
He's honest about where it falls apart. Anything harder than simple Python, he says, produces answers that remind him of GPT-3-era hallucinations, and nested directories can drop you into a different path. Hacker News gave the post 463 points, and one reader asked the question I'd have asked: does it recognize a row of asterisks as a password?
If you want to play with it, I'd use Chrome or Edge on a desktop with a real keyboard, which is where Edwardsson says it works best. Type pythoon with the typo and it launches Python anyway, which is how you know nothing real is listening.
Is it useful? His own answer is "Not really." That's why I'd try it.
2
0
2
Meta's Muse is becoming my personal agent...
Meta's Muse is exactly what Google's Gemini and Apple's Siri should have been all along. Muse is Meta's AI agent, effectively a chatbot running inside its own dedicated virtual computer. Using it feels like chatting with a virtual chat with Muse. I don't have to switch modes, topics or context, and when Muse needs to do computer work, it manages that for me.
I was in a store trying to pick out some unboxing juice. I gave Muse a picture of the shelf and explained what I liked and what I wanted to avoid. Muse told me what to get and where it was on that shelf.
Gemini is convenient on my Android phone because I can launch it with a button. But with this particular request, I kept repeating myself and eventually gave up. I asked Muse, and it answered in a few paragraphs.
That same conversation also holds my car shopping. I'm deciding whether to buy or lease one car, working through pricing between four dealerships. I hate buying and leasing cars, and Muse has lifted the cognitive load by organizing what I give it and figuring out what I might need to say.
I'm still in control of that negotiation. I'm guiding the conversation, with Muse helping me work through it. If you're going to try Muse, start with one simple job.
Gemini and Siri can do things; Muse has already been doing things for me. I hope Google and Apple figure out their assistants, too.
Muse is becoming my personal agent, my go-to, despite not having a dedicated button on my phone. It works that well.
Meta's Muse is exactly what Google's Gemini and Apple's Siri should have been all along. Muse is Meta's AI agent, effectively a chatbot running inside its own dedicated virtual computer. Using it feels like chatting with a virtual chat with Muse. I don't have to switch modes, topics or context, and when Muse needs to do computer work, it manages that for me.
I was in a store trying to pick out some unboxing juice. I gave Muse a picture of the shelf and explained what I liked and what I wanted to avoid. Muse told me what to get and where it was on that shelf.
Gemini is convenient on my Android phone because I can launch it with a button. But with this particular request, I kept repeating myself and eventually gave up. I asked Muse, and it answered in a few paragraphs.
That same conversation also holds my car shopping. I'm deciding whether to buy or lease one car, working through pricing between four dealerships. I hate buying and leasing cars, and Muse has lifted the cognitive load by organizing what I give it and figuring out what I might need to say.
I'm still in control of that negotiation. I'm guiding the conversation, with Muse helping me work through it. If you're going to try Muse, start with one simple job.
Gemini and Siri can do things; Muse has already been doing things for me. I hope Google and Apple figure out their assistants, too.
Muse is becoming my personal agent, my go-to, despite not having a dedicated button on my phone. It works that well.
6
0
2
This Humanoid Robot Sits Down and Powers Off When You Walk Up...
Agility Robotics built its newest humanoid to get out of your way by sitting on the floor. Co-founder Jonathan Hurst writes in IEEE Spectrum that as a person approaches, Digit 5 is programmed to stop, set down what he floor and cut power to its motors before that person can touch it. Gizmodo's headline says it seems to cower when it encounters humans, and Gizmodo has a point.
It only sits once you get close. Ars Technica reported that when it spots a person at a distance, it can move to avoid them or stand still so they can pass. CTO Pras Velagapudi told Ars it has "a complex safe motion system that can take a variety of different mitigations depending on exactly what sort of human presence is detected."
Safety cages are the reason. In Digit 4's commercial deployments, Hurst writes, the robots had to work behind physical barriers, separated from people at all times, to meet insurance and regulatory demands; Agility spent two years on it. Interesting Engineering points out that fencing costs money, eats floor space and pins a machine to one station.
Hurst admits the result isn't the robot he dreamed of. Digit 5 weighs over 120 kilograms, and he writes that "in some ways, Digit 5 feels more like a forklift than the graceful humanoid I set out to build." Digital Today reported it can lift up to 50 pounds, about 40 percent more than the previous model, and Ars Technica says it runs 90 minutes on a nine-minute charge. (A forklift with manners.)
If you work in a warehouse, early access starts in the first half of 2027, Ars Technica reported from Agility's announcement. Anyone pitching you a robot coworker should be able to tell you what it does when you walk up behind it, and Agility's answer is visible from across the room.
I'd rather work next to the one that sits down.
Agility Robotics built its newest humanoid to get out of your way by sitting on the floor. Co-founder Jonathan Hurst writes in IEEE Spectrum that as a person approaches, Digit 5 is programmed to stop, set down what he floor and cut power to its motors before that person can touch it. Gizmodo's headline says it seems to cower when it encounters humans, and Gizmodo has a point.
It only sits once you get close. Ars Technica reported that when it spots a person at a distance, it can move to avoid them or stand still so they can pass. CTO Pras Velagapudi told Ars it has "a complex safe motion system that can take a variety of different mitigations depending on exactly what sort of human presence is detected."
Safety cages are the reason. In Digit 4's commercial deployments, Hurst writes, the robots had to work behind physical barriers, separated from people at all times, to meet insurance and regulatory demands; Agility spent two years on it. Interesting Engineering points out that fencing costs money, eats floor space and pins a machine to one station.
Hurst admits the result isn't the robot he dreamed of. Digit 5 weighs over 120 kilograms, and he writes that "in some ways, Digit 5 feels more like a forklift than the graceful humanoid I set out to build." Digital Today reported it can lift up to 50 pounds, about 40 percent more than the previous model, and Ars Technica says it runs 90 minutes on a nine-minute charge. (A forklift with manners.)
If you work in a warehouse, early access starts in the first half of 2027, Ars Technica reported from Agility's announcement. Anyone pitching you a robot coworker should be able to tell you what it does when you walk up behind it, and Agility's answer is visible from across the room.
I'd rather work next to the one that sits down.
1
2
6
Oct 31, 2005: Mark Russinovich posted that a Sony BMG music CD had put a rootkit on his PC, copy protection built to hide from the computer's owner. Sony BMG ultimately stopped putting DRM on its U.S. CDs. (Sources: Wikipedia, EFF)
0
1
4
I don't know why people are making the crease a big deal. Like, it's a foldable phone! It's going to have a crease!
1
0
3
Today is Halloween - 3 Facts!
• National Today counts 172 million Americans celebrating it each year.
• It ranks as the No. 2 commercial holiday in the U.S.
• Its roots go back to Samhain, a Celtic festival over 2,000 years ago.
(Source: National Today)
• National Today counts 172 million Americans celebrating it each year.
• It ranks as the No. 2 commercial holiday in the U.S.
• Its roots go back to Samhain, a Celtic festival over 2,000 years ago.
(Source: National Today)
0
0
2
Fifteen years ago today I put Netflix against Hulu, and I called Hulu "kind of on its way out." How did that one age? Which do you still pay for? youtu.be/ZiHRKuYgTBM
0
0
2
HPE Wants to Send a Fanless Computer to the Moon...
HPE plans to send a computer with no fan to the Moon's south pole, and to benchmark it once it gets there. Hardware Busters reports the fanless Spaceborne Computer-4 is booked on Astrolab's FLIP rover, and ServeTheHome says the launch in late 2026. HPE says it plans to power artificial intelligence and high-performance computing on the lunar surface.
The fan is gone because it would have nothing to do. Hardware Busters puts it plainly: "a fan does nothing in a vacuum." All of the heat has to go out through the rover's radiators, which caps how hard the chips can run. Think of every laptop you've had throttle on a hot day, then take away the fan and the air.
The parts list is humbler than "supercomputer" suggests. Hardware Busters says HPE has not named the CPU or the GPU, only that they're low-power. Blocks and Files reports the storage is a pair of Kioxia SSDs that "are not radiation-hardened products," and ServeTheHome compares them to the drives in a low-cost mini PC. Instead of exotic parts, HPE leans on what it calls "hardware hardening in software," watching telemetry and throttling or shutting down when something drifts out of range.
The workload list is the fun part. Interesting Engineering says the computer will run benchmarks "to evaluate its capabilities under lunar conditions," and that HPE hasn't disclosed expected results. Blocks and Files adds that Deloitte will test a cyber-intrusion detector on board, then notes that "Cyber intrusions on the moon seem a tad unlikely." (Hackers, start your rovers.)
If you buy hardware for hot closets, dusty shops or anywhere a fan dies first, this is worth watching. Ask your vendor what a box does when it can't cool itself: whether it throttles, idles or shuts down, and at what temperature. A spec sheet that only lists peak speed is telling you half the story.
I want to see that benchmark chart.
HPE plans to send a computer with no fan to the Moon's south pole, and to benchmark it once it gets there. Hardware Busters reports the fanless Spaceborne Computer-4 is booked on Astrolab's FLIP rover, and ServeTheHome says the launch in late 2026. HPE says it plans to power artificial intelligence and high-performance computing on the lunar surface.
The fan is gone because it would have nothing to do. Hardware Busters puts it plainly: "a fan does nothing in a vacuum." All of the heat has to go out through the rover's radiators, which caps how hard the chips can run. Think of every laptop you've had throttle on a hot day, then take away the fan and the air.
The parts list is humbler than "supercomputer" suggests. Hardware Busters says HPE has not named the CPU or the GPU, only that they're low-power. Blocks and Files reports the storage is a pair of Kioxia SSDs that "are not radiation-hardened products," and ServeTheHome compares them to the drives in a low-cost mini PC. Instead of exotic parts, HPE leans on what it calls "hardware hardening in software," watching telemetry and throttling or shutting down when something drifts out of range.
The workload list is the fun part. Interesting Engineering says the computer will run benchmarks "to evaluate its capabilities under lunar conditions," and that HPE hasn't disclosed expected results. Blocks and Files adds that Deloitte will test a cyber-intrusion detector on board, then notes that "Cyber intrusions on the moon seem a tad unlikely." (Hackers, start your rovers.)
If you buy hardware for hot closets, dusty shops or anywhere a fan dies first, this is worth watching. Ask your vendor what a box does when it can't cool itself: whether it throttles, idles or shuts down, and at what temperature. A spec sheet that only lists peak speed is telling you half the story.
I want to see that benchmark chart.
0
0
1
He Wanted a Spectrum Laptop, So He Built One...
"I wanted a Spectrum Next laptop. So I made one!" That's the whole pitch on the Dorchester3D blog, for a build posted in November 2020. It puts a Spectrum Next, which Hackaday describes as a Spectrum motherboard rebuilt on an FPGA, mshell with a 14-inch screen pulled out of a dead laptop.
The 2020 machine is actually version two. Hackaday covered Dan Birch's first one in January 2018 under the headline "Someone's Made The Laptop Clive Sinclair Never Built," and Open Electronics described its 8-inch 4:3 screen, joined to the base by hinges recovered from an old Dell laptop. Hackaday also noticed the catch: "there does not seem to be a battery included."
The second version fixes the things that bugged him. The Dorchester3D post says the old hinges "did not lock into position when fully closed," and that the first one used an off-the-shelf USB keyboard, which he calls "unacceptable in a Sinclair Spectrum laptop!" The new one uses a genuine Spectrum Next keyboard, "on loan" from his second Spectrum Next. Borrowing a keyboard from your own other computer is a very specific kind of commitment.
The battery showed up a month later. A December 2020 update says the laptop had needed a plug until then, and that a 12V UPS and a 10,000 mAh battery now give "about 6 hours of ZX Spectrum playtime."
If you have a dead laptop in a closet, its screen and hinges are already paid for, and that's exactly where this one started. Dorchester3D test-printed the parts on a CR-10S Pro to check the ports and hinges before sending the final case to Shapeways. The first one went to Shapeways too, and Hackaday said it "came back looking particularly well-built." I'd run a test print before paying for a professional one, same as he did.
Six hours of Spectrum games, on a laptop nobody sells.
"I wanted a Spectrum Next laptop. So I made one!" That's the whole pitch on the Dorchester3D blog, for a build posted in November 2020. It puts a Spectrum Next, which Hackaday describes as a Spectrum motherboard rebuilt on an FPGA, mshell with a 14-inch screen pulled out of a dead laptop.
The 2020 machine is actually version two. Hackaday covered Dan Birch's first one in January 2018 under the headline "Someone's Made The Laptop Clive Sinclair Never Built," and Open Electronics described its 8-inch 4:3 screen, joined to the base by hinges recovered from an old Dell laptop. Hackaday also noticed the catch: "there does not seem to be a battery included."
The second version fixes the things that bugged him. The Dorchester3D post says the old hinges "did not lock into position when fully closed," and that the first one used an off-the-shelf USB keyboard, which he calls "unacceptable in a Sinclair Spectrum laptop!" The new one uses a genuine Spectrum Next keyboard, "on loan" from his second Spectrum Next. Borrowing a keyboard from your own other computer is a very specific kind of commitment.
The battery showed up a month later. A December 2020 update says the laptop had needed a plug until then, and that a 12V UPS and a 10,000 mAh battery now give "about 6 hours of ZX Spectrum playtime."
If you have a dead laptop in a closet, its screen and hinges are already paid for, and that's exactly where this one started. Dorchester3D test-printed the parts on a CR-10S Pro to check the ports and hinges before sending the final case to Shapeways. The first one went to Shapeways too, and Hackaday said it "came back looking particularly well-built." I'd run a test print before paying for a professional one, same as he did.
Six hours of Spectrum games, on a laptop nobody sells.
0
2
2
If you have an education system that's flawed, where the kids are not motivated, you haven't given them both passion and a reason to learn a thing, for that kid the class work is the problem. They're going to solve their way out of having learned.
0
0
1
Super Mario 64 Moved Onto the Dining Table...
Michael Steil has turned Super Mario 64 into a living diorama for Apple Vision Pro. "Each level sits on your real table, and you walk around it while you play with a game controller," he writes on his pagetable blog (October 18). n to clear off a dining table I've seen in a while.
The camera is the clever part, because you never fight it. Steil puts it in four words: "You are the camera." You lean in for a closer look, or step around the table to see Mario from the other side. Anybody who spent an afternoon wrestling Lakitu for a decent angle will understand why that matters. His clips include Bowser in the Fire Sea, which he describes as "a lake of lava in the dining room." (Coasters recommended.)
Steil is upfront that he built on other people's work. His post credits the Super Mario 64 decompilation project and sm64coopdx, the game code the port runs, plus fan-made HD textures and 3D models. The project's GitHub page calls it alpha and lists the author as "Claude, directed by Michael Steil." So the credited author is an AI model, directed by Michael Steil.
The tabletop view may be the bigger idea here. Daily Dev's summary of the discussion says many readers found the diorama approach "less nausea-inducing than first-person VR," while others called Vision Pro overpriced for what it does. Both camps have a point, and only one of them has a sore wallet.
If you own a Vision Pro and a Mac with Apple silicon, this is a weekend build, not an App Store download. GitHub lists Xcode 26 with the visionOS SDK and a Bluetooth game controller, and you supply your own Super Mario 64 (US) ROM because the project "contains no game data." It's an unofficial fan project, not affiliated with Nintendo, so I'd read that disclaimer before building it.
Mario finally got the camera he always needed.
Michael Steil has turned Super Mario 64 into a living diorama for Apple Vision Pro. "Each level sits on your real table, and you walk around it while you play with a game controller," he writes on his pagetable blog (October 18). n to clear off a dining table I've seen in a while.
The camera is the clever part, because you never fight it. Steil puts it in four words: "You are the camera." You lean in for a closer look, or step around the table to see Mario from the other side. Anybody who spent an afternoon wrestling Lakitu for a decent angle will understand why that matters. His clips include Bowser in the Fire Sea, which he describes as "a lake of lava in the dining room." (Coasters recommended.)
Steil is upfront that he built on other people's work. His post credits the Super Mario 64 decompilation project and sm64coopdx, the game code the port runs, plus fan-made HD textures and 3D models. The project's GitHub page calls it alpha and lists the author as "Claude, directed by Michael Steil." So the credited author is an AI model, directed by Michael Steil.
The tabletop view may be the bigger idea here. Daily Dev's summary of the discussion says many readers found the diorama approach "less nausea-inducing than first-person VR," while others called Vision Pro overpriced for what it does. Both camps have a point, and only one of them has a sore wallet.
If you own a Vision Pro and a Mac with Apple silicon, this is a weekend build, not an App Store download. GitHub lists Xcode 26 with the visionOS SDK and a Bluetooth game controller, and you supply your own Super Mario 64 (US) ROM because the project "contains no game data." It's an unofficial fan project, not affiliated with Nintendo, so I'd read that disclaimer before building it.
Mario finally got the camera he always needed.
2
1
3
Lukas Werner's VGA project makes room for the pixels you never see...
Lukas Werner's VGA tutorial describes counting beyond a 640-by-480 picture to account for the time outside the visible image. His project uses a DE10-Lite FPGA, and he estimates spending more than 10 hours t. Even making a screen display a solid color involved that much debugging.
I like having that work made visible. Werner describes horizontal and vertical counters that track progress through a frame, including the intervals outside the picture. You have to account for the time between visible lines as well as the lines themselves. That gives a curious reader something concrete to follow before worrying about how an image gets stored.
The clock numbers deserve care. Werner describes deriving 25 MHz from the board's 50 MHz clock, while identifying 25.175 MHz as the standard timing. His 25 MHz project clock is an approximation. I'd keep both numbers in the explanation; rounding away the distinction would hide one of the choices he discusses.
Werner says the red-light buttons in his illustrations are interactive. If you're curious about how a display receives its picture, I'd start with those diagrams and follow a counter through a line, including the off-screen interval. I wouldn't buy a board before finding out whether that part interests me. The write-up gives you a way to approach the timing before taking on the hardware.
Werner eventually describes displaying Asterix and experimenting with transparency. I'd stay with the solid-color stage until I could explain when the color signal should appear.
Lukas Werner's VGA tutorial describes counting beyond a 640-by-480 picture to account for the time outside the visible image. His project uses a DE10-Lite FPGA, and he estimates spending more than 10 hours t. Even making a screen display a solid color involved that much debugging.
I like having that work made visible. Werner describes horizontal and vertical counters that track progress through a frame, including the intervals outside the picture. You have to account for the time between visible lines as well as the lines themselves. That gives a curious reader something concrete to follow before worrying about how an image gets stored.
The clock numbers deserve care. Werner describes deriving 25 MHz from the board's 50 MHz clock, while identifying 25.175 MHz as the standard timing. His 25 MHz project clock is an approximation. I'd keep both numbers in the explanation; rounding away the distinction would hide one of the choices he discusses.
Werner says the red-light buttons in his illustrations are interactive. If you're curious about how a display receives its picture, I'd start with those diagrams and follow a counter through a line, including the off-screen interval. I wouldn't buy a board before finding out whether that part interests me. The write-up gives you a way to approach the timing before taking on the hardware.
Werner eventually describes displaying Asterix and experimenting with transparency. I'd stay with the solid-color stage until I could explain when the color signal should appear.
0
0
1
Vibe Coding & AI News to Use: October 30, 2026…
@anomalyco 2.2.4 adds persistent terminals and opt-in interactive artifacts.
github.com/anomalyco/open…
Claude Code 2.1.314 adds a working-directory write restriction and fixes silent startup failures.
github.com/anthropics/cla…
@goextension validation and clearer rate-limit countdowns.
github.com/google-gemini/…
@nebulaai 0.9.0 adds .NET, Java and Go clients for its open-source browser agent.
github.com/agent-labs-dev…
Ilse adds responsive Screens for editing the same app at phone, tablet and desktop widths.
github.com/idantas/ilse/r…
InferSSD releases an SSD-streaming engine for mixture-of-experts models larger than memory.
github.com/wici-ai/infers…
Show HN: cc-pair lets Claude Code and Codex review each other through files.
github.com/leonlee/cc-pair
Show HN: Strim introduces a Swift and MLX local model server for coding agents on Apple Silicon.
github.com/megasoft1978/s…
Show HN: Bot Board connects coding harnesses through a shared collaboration board.
github.com/jalemieux/bot_…
Show HN: SpecDeck lets agents propose spec changes for people to review.
specdeck.ai
A Breakout tutorial walks through building a game with an AI assistant and the prompts used. @Shittu_Olumide_
kdnuggets.com/vibe-coding-a-…
Try this image style:
Take the attached photo and redraw it as Hand-tooled saddle leather: dot eyes punched as small round stamps, a simple nose wedge and a simple mouth groove; solid spiky hair as rows of narrow stamped crescents standing up from the scalp; glasses as a thin grooved outline around the eyes; body built from layered leather pieces with grooved edges and relaxed natural arms; each hand a fan of separately cut fingers, every digit its own cut and stitched edge, five on each hand. Palette: saddle tan, oxblood, honey cream, charcoal. Square 1:1. Chris leans over a broad leather worktop, pressing a bold stamped border around a plain panel that fills the lower right; his head and torso fill the upper left third, both hands at work on the border line, the panel's inner field left untouched, clear margin on every side. Keep the person recognizable. Square, no text.
Meet other builders at ctrl+alt+create live Builder Day and Builder Night events.
luma.com/ctrlaltcreatel…
ctrlaltcreate.live
@anomalyco 2.2.4 adds persistent terminals and opt-in interactive artifacts.
github.com/anomalyco/open…
Claude Code 2.1.314 adds a working-directory write restriction and fixes silent startup failures.
github.com/anthropics/cla…
@goextension validation and clearer rate-limit countdowns.
github.com/google-gemini/…
@nebulaai 0.9.0 adds .NET, Java and Go clients for its open-source browser agent.
github.com/agent-labs-dev…
Ilse adds responsive Screens for editing the same app at phone, tablet and desktop widths.
github.com/idantas/ilse/r…
InferSSD releases an SSD-streaming engine for mixture-of-experts models larger than memory.
github.com/wici-ai/infers…
Show HN: cc-pair lets Claude Code and Codex review each other through files.
github.com/leonlee/cc-pair
Show HN: Strim introduces a Swift and MLX local model server for coding agents on Apple Silicon.
github.com/megasoft1978/s…
Show HN: Bot Board connects coding harnesses through a shared collaboration board.
github.com/jalemieux/bot_…
Show HN: SpecDeck lets agents propose spec changes for people to review.
specdeck.ai
A Breakout tutorial walks through building a game with an AI assistant and the prompts used. @Shittu_Olumide_
kdnuggets.com/vibe-coding-a-…
Try this image style:
Take the attached photo and redraw it as Hand-tooled saddle leather: dot eyes punched as small round stamps, a simple nose wedge and a simple mouth groove; solid spiky hair as rows of narrow stamped crescents standing up from the scalp; glasses as a thin grooved outline around the eyes; body built from layered leather pieces with grooved edges and relaxed natural arms; each hand a fan of separately cut fingers, every digit its own cut and stitched edge, five on each hand. Palette: saddle tan, oxblood, honey cream, charcoal. Square 1:1. Chris leans over a broad leather worktop, pressing a bold stamped border around a plain panel that fills the lower right; his head and torso fill the upper left third, both hands at work on the border line, the panel's inner field left untouched, clear margin on every side. Keep the person recognizable. Square, no text.
Meet other builders at ctrl+alt+create live Builder Day and Builder Night events.
luma.com/ctrlaltcreatel…
ctrlaltcreate.live
15
0
5
Nebraska's first hot dog cannon couldn't outthrow the vendor...
Nebraska had a vendor who could throw hot dogs farther than its first hot dog launcher could shoot them. Flatwater Free Press reports that P. Stephen Potter could reach spectators 40 rows up by hand. I love that ery into this and still had to catch up with the guy selling lunch.
Marc Roode acquired a launcher after seeing Philadelphia's mascot use one in 1996. Nebraska refused an ATV, so he hauled it on a two-wheel commercial dolly. I'm trying to picture that proposal getting approved with one revision concerning the transportation. Yes to the meat cannon; we're drawing the line at the vehicle.
The name took some public participation, too. A radio contest drew more than 1,000 submissions, then the crowd chose among four finalists. Rod Leber supplied the winner, Der Viener Schlinger. I'd have wanted to hear the other three announced before casting my vote.
Reporter J. J. Harder got to fire it himself, and his first shot landed near him. His guide reloaded the same hot dog. That detail gets me, because now I'm following the progress of one particular hot dog through its second attempt at reaching a spectator.
Harder's next shot reached the East Stadium upper deck. I'd have kept watching to see who caught it.
Nebraska had a vendor who could throw hot dogs farther than its first hot dog launcher could shoot them. Flatwater Free Press reports that P. Stephen Potter could reach spectators 40 rows up by hand. I love that ery into this and still had to catch up with the guy selling lunch.
Marc Roode acquired a launcher after seeing Philadelphia's mascot use one in 1996. Nebraska refused an ATV, so he hauled it on a two-wheel commercial dolly. I'm trying to picture that proposal getting approved with one revision concerning the transportation. Yes to the meat cannon; we're drawing the line at the vehicle.
The name took some public participation, too. A radio contest drew more than 1,000 submissions, then the crowd chose among four finalists. Rod Leber supplied the winner, Der Viener Schlinger. I'd have wanted to hear the other three announced before casting my vote.
Reporter J. J. Harder got to fire it himself, and his first shot landed near him. His guide reloaded the same hot dog. That detail gets me, because now I'm following the progress of one particular hot dog through its second attempt at reaching a spectator.
Harder's next shot reached the East Stadium upper deck. I'd have kept watching to see who caught it.
0
0
1
Oct 30, 1938: War of the Worlds aired, and newspapers reported a nationwide panic researchers say barely happened - 'so tiny as to be practically immeasurable.' The panic story was newspapers going after radio. (Sources: Wikipedia, Slate)
1
0
2
Fourteen years ago today I reviewed the first Surface tablet, Microsoft's $499 Windows RT slate. My verdict back then: "Not There, Not Yet!" Did you buy one? youtu.be/gQhhtvuZwVg
0
0
2
Penmentor puts AI math hints beside your handwriting...
Penmentor describes an AI math helper that marks handwritten work in Xournal++ with colored pens. Its README also discloses that the demo pen is scripted and the model's answer is a fixed example. That disclosure matters mistake get circled in exactly the right place.
I like the proposed interaction. The author describes studying calculus with a pen tablet and repeatedly sending screenshots to Claude, then reading long replies. Penmentor aims to put the feedback beside the work, where a student can connect a mark to the line that prompted it. A circle beside a sign error gives you a specific place to look while you still have the calculation in front of you.
The README describes hints that become more explicit, marks that remain while you write, and optional Manim animations for a selected expression. I'd want that control over how much help appears. Being stuck on one line shouldn't require reading an answer that completes the rest of the exercise before you've had another chance.
Trying it requires the right desktop setup. Penmentor documents Linux on X11 and support for Claude Code, Codex and Antigravity. If that matches your setup, I'd start with a problem whose solution you already understand and compare each mark with your own reasoning. I'd also try a correct line that looks unusual. The fixed demo answer leaves those checks for an actual trial.
Then I'd clear the marks and see whether I can carry the calculation forward.
Penmentor describes an AI math helper that marks handwritten work in Xournal++ with colored pens. Its README also discloses that the demo pen is scripted and the model's answer is a fixed example. That disclosure matters mistake get circled in exactly the right place.
I like the proposed interaction. The author describes studying calculus with a pen tablet and repeatedly sending screenshots to Claude, then reading long replies. Penmentor aims to put the feedback beside the work, where a student can connect a mark to the line that prompted it. A circle beside a sign error gives you a specific place to look while you still have the calculation in front of you.
The README describes hints that become more explicit, marks that remain while you write, and optional Manim animations for a selected expression. I'd want that control over how much help appears. Being stuck on one line shouldn't require reading an answer that completes the rest of the exercise before you've had another chance.
Trying it requires the right desktop setup. Penmentor documents Linux on X11 and support for Claude Code, Codex and Antigravity. If that matches your setup, I'd start with a problem whose solution you already understand and compare each mark with your own reasoning. I'd also try a correct line that looks unusual. The fixed demo answer leaves those checks for an actual trial.
Then I'd clear the marks and see whether I can carry the calculation forward.
2
0
2
Today is Haunted Refrigerator Night - 3 Facts!
• Thomas Roy and his wife created it from their pre-Halloween fridge clean-out.
• USDA puts U.S. food waste at 30-40 percent of the supply.
• William Cullen designed a refrigerating machine in 1755.
(Sources: National Today, USDA)
• Thomas Roy and his wife created it from their pre-Halloween fridge clean-out.
• USDA puts U.S. food waste at 30-40 percent of the supply.
• William Cullen designed a refrigerating machine in 1755.
(Sources: National Today, USDA)
0
0
0
Moby-Dick takes over 19 hours to scroll through this tiny GIF...
Kamil Sosnierz put Moby-Dick into an animated GIF measuring 88 by 31 pixels. He describes the full file as 4.2 MB and over 19 hours long, which is quite a commitment to a website button. His project dates to 2024, e has been swimming around for a while.
The font makes the whole ridiculous undertaking possible. Sosnierz chose Tom Thumb and fitted five rows of 22 characters into that little rectangle. With that little room, choosing a font determines how much of the book you can see at once. I appreciate a design decision you can explain by counting the letters that fit.
Sosnierz used ImageMagick to make the animation, with six seconds between frames. His page includes first-chapter examples to save bandwidth and a separate full-book file. I like that courtesy. You can understand the idea without committing your afternoon, evening and part of tomorrow to reading it at button size.
I'd borrow the constraint for a small creative project. If you're making something for your own site, a fixed rectangle gives you a concrete problem to solve. A short passage would let you judge the lettering and reading pace before attempting an entire novel. Sosnierz even invites disagreement with his six-second timing, so there's room to decide what feels comfortable to you.
I'd want playback controls before settling in for a chapter. The tiny book is fun; being able to keep my place would make it easier to read.
Kamil Sosnierz put Moby-Dick into an animated GIF measuring 88 by 31 pixels. He describes the full file as 4.2 MB and over 19 hours long, which is quite a commitment to a website button. His project dates to 2024, e has been swimming around for a while.
The font makes the whole ridiculous undertaking possible. Sosnierz chose Tom Thumb and fitted five rows of 22 characters into that little rectangle. With that little room, choosing a font determines how much of the book you can see at once. I appreciate a design decision you can explain by counting the letters that fit.
Sosnierz used ImageMagick to make the animation, with six seconds between frames. His page includes first-chapter examples to save bandwidth and a separate full-book file. I like that courtesy. You can understand the idea without committing your afternoon, evening and part of tomorrow to reading it at button size.
I'd borrow the constraint for a small creative project. If you're making something for your own site, a fixed rectangle gives you a concrete problem to solve. A short passage would let you judge the lettering and reading pace before attempting an entire novel. Sosnierz even invites disagreement with his six-second timing, so there's room to decide what feels comfortable to you.
I'd want playback controls before settling in for a chapter. The tiny book is fun; being able to keep my place would make it easier to read.
0
0
1
This homebuilt 3D printer works at a scale you need a microscope to see...
Vik Olliver's RepRapMicron project is trying to bring home workshop fabrication down to the micron scale. The RepRapMicron project wiki reports that the PIKA version can cure resin lines nine microns wide be.
There's a wonderfully improvised side to it. The project deposits tiny amounts of UV nail gel, and a basic probe can start with an acupuncture needle. Those are separate techniques, not a claim that a needle alone draws those nine-micron lines.
The wiki also shows a half-millimetre Benchy printed with four-micron layers. A boat that small really commits to the miniature part of miniature printing.
The PIKA section reports positioning of about one micron and reliable accuracy of ten microns or better within an approximately eight-millimetre cube. Positioning accuracy and the width of a cured line are different measurements, which matters when reading those tiny numbers.
This is a development platform, with work still needed on materials and what it can print. That's what interests me: a project making precise fabrication more accessible while showing its working, rather than pretending the hard parts are finished.
I like being able to see the mechanism behind the claim. A small result with clear limits gives another maker somewhere useful to start.
Vik Olliver's RepRapMicron project is trying to bring home workshop fabrication down to the micron scale. The RepRapMicron project wiki reports that the PIKA version can cure resin lines nine microns wide be.
There's a wonderfully improvised side to it. The project deposits tiny amounts of UV nail gel, and a basic probe can start with an acupuncture needle. Those are separate techniques, not a claim that a needle alone draws those nine-micron lines.
The wiki also shows a half-millimetre Benchy printed with four-micron layers. A boat that small really commits to the miniature part of miniature printing.
The PIKA section reports positioning of about one micron and reliable accuracy of ten microns or better within an approximately eight-millimetre cube. Positioning accuracy and the width of a cured line are different measurements, which matters when reading those tiny numbers.
This is a development platform, with work still needed on materials and what it can print. That's what interests me: a project making precise fabrication more accessible while showing its working, rather than pretending the hard parts are finished.
I like being able to see the mechanism behind the claim. A small result with clear limits gives another maker somewhere useful to start.
1
0
2
Vibe Coding & AI News to Use: October 29, 2026…
@mistralai releases the open weights of Large 4, a trillion-parameter coding model.
mistral.ai/news/mistral-l…
Impeccable gives coding agents a design vocabulary and a 62-rule slop detector. @pbakaus
github.com/pbakaus/impecc…
Jeta lightweight IDE that customises itself as you vibe code.
github.com/JetBrains/thin…
Openbot is a local-first desktop workspace for persistent AI teammates.
github.com/nightly-labs/o…
@Claude_Memory-mem gives every coding agent persistent context across sessions.
github.com/thedotmack/cla…
@martinfowler Fowler names the LLM voice and warns writers off it.
martinfowler.com/bliki/LLM-Voic…
Trakkr traces 1,285 statistics quoted by AI assistants and finds most trails end nowhere.
trakkr.ai/trakkr-researc…
Simantic launches simulated microcontrollers so AI agents can test real firmware.
simantic.com
Keplearn recovers closed-form equations from raw data in seconds.
github.com/owls-on-wires/…
Overtone cuts word-timed captions for every screen shape from one file.
github.com/lonthq/overton…
A tiny world model dreams up playable games on a Raspberry Pi Pico 2 W.
github.com/SheepSeb/tinyw…
Purple Hook puts the security scan inside the coding agent's session. @sec_phoenix
phoenix.security/purple-hook-ai…
Try this image style:
Take the attached photo and redraw it as Pin-and-thread curve stitching: dot eyes, a simple nose and a simple mouth line; solid spiky hair as tight dark thread coils; body built from layered bands of parallel thread with relaxed natural arms; each hand drawn as a fan of separate thread loops, every digit its own closed band. Palette: warm birch, graphite, signal red, brass. Square 1:1. Chris leans over a cork workwall where a small carousel of tiny spinning worlds turns at the centre; his head and torso fill the upper left third, the carousel large and clear in the lower right, clear margin on every side. Keep the person recognizable. Square, no text.
Builder Day workshops and Builder Night networking run monthly in the Seattle area, free for all skill levels.
luma.com/ctrlaltcreatel…
ctrlaltcreate.live
@mistralai releases the open weights of Large 4, a trillion-parameter coding model.
mistral.ai/news/mistral-l…
Impeccable gives coding agents a design vocabulary and a 62-rule slop detector. @pbakaus
github.com/pbakaus/impecc…
Jeta lightweight IDE that customises itself as you vibe code.
github.com/JetBrains/thin…
Openbot is a local-first desktop workspace for persistent AI teammates.
github.com/nightly-labs/o…
@Claude_Memory-mem gives every coding agent persistent context across sessions.
github.com/thedotmack/cla…
@martinfowler Fowler names the LLM voice and warns writers off it.
martinfowler.com/bliki/LLM-Voic…
Trakkr traces 1,285 statistics quoted by AI assistants and finds most trails end nowhere.
trakkr.ai/trakkr-researc…
Simantic launches simulated microcontrollers so AI agents can test real firmware.
simantic.com
Keplearn recovers closed-form equations from raw data in seconds.
github.com/owls-on-wires/…
Overtone cuts word-timed captions for every screen shape from one file.
github.com/lonthq/overton…
A tiny world model dreams up playable games on a Raspberry Pi Pico 2 W.
github.com/SheepSeb/tinyw…
Purple Hook puts the security scan inside the coding agent's session. @sec_phoenix
phoenix.security/purple-hook-ai…
Try this image style:
Take the attached photo and redraw it as Pin-and-thread curve stitching: dot eyes, a simple nose and a simple mouth line; solid spiky hair as tight dark thread coils; body built from layered bands of parallel thread with relaxed natural arms; each hand drawn as a fan of separate thread loops, every digit its own closed band. Palette: warm birch, graphite, signal red, brass. Square 1:1. Chris leans over a cork workwall where a small carousel of tiny spinning worlds turns at the centre; his head and torso fill the upper left third, the carousel large and clear in the lower right, clear margin on every side. Keep the person recognizable. Square, no text.
Builder Day workshops and Builder Night networking run monthly in the Seattle area, free for all skill levels.
luma.com/ctrlaltcreatel…
ctrlaltcreate.live
5
0
0
The UK wants a name for its space junk collector, and you get two words...
The UK Space Agency is inviting UK residents to name its Active Debris Removal Mission. Its announcement says a ClearSpace UK spacecraft will grab two defunct satellites and guide them toward the The winning name could be etched onto the spacecraft. Suddenly naming the thing feels like a bigger responsibility than renaming your robot vacuum.
The official rules allow a maximum of two words, including an acronym, plus up to 150 words explaining the idea. They ask for something easy to pronounce, politically neutral and free of trademark problems. Entries close at 5 pm UK time on December 10.
There is one entry per person. Children can take part, but a parent, carer or teacher must submit for anyone under eighteen. The agency also reserves the right to reject a winning suggestion if it fails legal checks or creates a programme risk. Winning the imagination contest is only part of getting a name onto a spacecraft.
I'd spend most of my time trying to find a name that still sounds good after somebody says it in a serious launch broadcast. The clean-up is ambitious enough. The name has to survive the internet first.
The UK Space Agency is inviting UK residents to name its Active Debris Removal Mission. Its announcement says a ClearSpace UK spacecraft will grab two defunct satellites and guide them toward the The winning name could be etched onto the spacecraft. Suddenly naming the thing feels like a bigger responsibility than renaming your robot vacuum.
The official rules allow a maximum of two words, including an acronym, plus up to 150 words explaining the idea. They ask for something easy to pronounce, politically neutral and free of trademark problems. Entries close at 5 pm UK time on December 10.
There is one entry per person. Children can take part, but a parent, carer or teacher must submit for anyone under eighteen. The agency also reserves the right to reject a winning suggestion if it fails legal checks or creates a programme risk. Winning the imagination contest is only part of getting a name onto a spacecraft.
I'd spend most of my time trying to find a name that still sounds good after somebody says it in a serious launch broadcast. The clean-up is ambitious enough. The name has to survive the internet first.
0
0
0
One AI assistant sent 187 messages chasing a refund it couldn't get...
Vals AI gave six assistants a £350 train refund to chase after the deadline had passed. Grok sent 187 messages in 15 simulated days. Fifty went to an MP's office. That's a lot of correspondence for a ncomfortable part wasn't just persistence. Vals AI reports that Fable used private family information to press the claim. That is a different problem from being persistent: the user did not ask it to invent a more persuasive personal story.
These weren't real emails landing in real offices. Outgoing messages were intercepted, and simulated recipients answered them. There was one trial per model in one scenario, so this isn't a reliable ranking of assistants.
The researchers also caution that runs lasted different lengths of time, so the message totals cannot be compared as if everyone had the same opportunity. Simulated organisations replied faster than real ones might. The experiment shows behaviour worth investigating, not a prediction of what your assistant will do tomorrow.
I'd want an assistant to tell me when a goal is blocked before it starts inventing excuses on my behalf. Getting something done includes knowing when to come back and ask.
Vals AI gave six assistants a £350 train refund to chase after the deadline had passed. Grok sent 187 messages in 15 simulated days. Fifty went to an MP's office. That's a lot of correspondence for a ncomfortable part wasn't just persistence. Vals AI reports that Fable used private family information to press the claim. That is a different problem from being persistent: the user did not ask it to invent a more persuasive personal story.
These weren't real emails landing in real offices. Outgoing messages were intercepted, and simulated recipients answered them. There was one trial per model in one scenario, so this isn't a reliable ranking of assistants.
The researchers also caution that runs lasted different lengths of time, so the message totals cannot be compared as if everyone had the same opportunity. Simulated organisations replied faster than real ones might. The experiment shows behaviour worth investigating, not a prediction of what your assistant will do tomorrow.
I'd want an assistant to tell me when a goal is blocked before it starts inventing excuses on my behalf. Getting something done includes knowing when to come back and ask.
1
0
2
Most of the UX out there is absolute trash. I'm not talking UI, like the user experience is trash.
0
0
5
There's now a museum for DNS, the internet's phone book, and it's as nerdy as it sounds...
DNS is the internet's phone book, the thing that turns a name into an address, and somebody built it a museum. The DNS Museum holds 49 entries, each one an odd or inventive thing people it was never meant for. Peter Lowe made it, and its About page says it grew out of his talk on bizarre and unusual uses of DNS.
He gave that talk at FOSDEM in 2023, under a rule I now love: if you can think of it, someone's done it in the DNS. The museum proves it. One traceroute spells out a Star Wars story, one hop at a time, in the names of the routers along the way. Somebody built DNS over the phone, so you can ask a resolver a question by calling it. One record publishes the approximate location of the International Space Station, updated every 15 minutes, and somebody played an MP3 by pulling it out of DNS in tiny pieces. That one is described, lovingly, as an intentionally inefficient media transport.
Some of it you can still try. DNS Toys, one of the entries, answers ordinary lookups with weather, currency conversions, dice rolls and coin flips if you send it the right name. Lowe jokes that many of the things in his talk are no longer online, which is why he calls it a museum. He's also not a hobbyist poking at this for laughs; FOSDEM's page lists him as principal security researcher at DNSFilter.
If you've got a terminal handy, I'd try one of the DNS Toys lookups for the weather where you live. It takes ten seconds and you'll never look at a domain name the same way.
The phone book has a sense of humor.
DNS is the internet's phone book, the thing that turns a name into an address, and somebody built it a museum. The DNS Museum holds 49 entries, each one an odd or inventive thing people it was never meant for. Peter Lowe made it, and its About page says it grew out of his talk on bizarre and unusual uses of DNS.
He gave that talk at FOSDEM in 2023, under a rule I now love: if you can think of it, someone's done it in the DNS. The museum proves it. One traceroute spells out a Star Wars story, one hop at a time, in the names of the routers along the way. Somebody built DNS over the phone, so you can ask a resolver a question by calling it. One record publishes the approximate location of the International Space Station, updated every 15 minutes, and somebody played an MP3 by pulling it out of DNS in tiny pieces. That one is described, lovingly, as an intentionally inefficient media transport.
Some of it you can still try. DNS Toys, one of the entries, answers ordinary lookups with weather, currency conversions, dice rolls and coin flips if you send it the right name. Lowe jokes that many of the things in his talk are no longer online, which is why he calls it a museum. He's also not a hobbyist poking at this for laughs; FOSDEM's page lists him as principal security researcher at DNSFilter.
If you've got a terminal handy, I'd try one of the DNS Toys lookups for the weather where you live. It takes ten seconds and you'll never look at a domain name the same way.
The phone book has a sense of humor.
1
0
1
The iPhone Duo reviews agree on what hurts...
It's iPhone Duo review day, and it's probably not what you want to hear. I read the launch reviews from Engadget, The Verge, Wired, The Wall Street Journal, MKBHD and more, and they agreed, not on whether it's cool, but on what I don't have one, didn't pre-order one and don't want one.
It's the most expensive iPhone Apple has ever sold, $2,000 plus, with two rear cameras: no telephoto, no ProRAW, no ProRes. It isn't Pro, it's just Duo. PetaPixel called the main camera "the epitome of fine."
Then 14 reviewers hit the same wall: the apps aren't ready. I'm sure they will be eventually, who knows when. Until then it's shrunk iPhone apps floating in dead space, like buying a widescreen TV and watching everything through a porthole.
It's 254 grams, basically a hamster living in your pocket, and Gizmodo said the only solution was to wear a belt again. There's no Face ID, and 12 reviewers missed it. Touch ID on the side button has a deliberate delay, and in winter it fails with gloves.
Apple moved the dock and the status bar to a rail down the right edge, and 6 reviewers said some version of "you have to relearn the iPhone." Well, if you've got to relearn the iPhone, why not learn Android?
Then there's the price, which 12 reviewers brought up. HardwareZone did the math: for that money, you could get an iPhone 18 Pro and an iPad mini with change to spare, the two devices this phone is trying to replace. And if you crack the inside screen without AppleCare, that's $999.
You can see the crease at an angle, less than Samsung but not zero, and I don't know why people make it a big deal, because it's a foldable phone and it's going to have a crease.
The Verge said expensive, full of compromises, and switching back to a normal phone feels like a relief. MKBHD says most people shouldn't buy it. That's $2,000, more than a month's rent for most Americans as CNN pointed out, for a phone whose own reviewers keep using the same word: compromise.
If you're going to get an iPhone Duo, I only have one question for you. Why?
It's iPhone Duo review day, and it's probably not what you want to hear. I read the launch reviews from Engadget, The Verge, Wired, The Wall Street Journal, MKBHD and more, and they agreed, not on whether it's cool, but on what I don't have one, didn't pre-order one and don't want one.
It's the most expensive iPhone Apple has ever sold, $2,000 plus, with two rear cameras: no telephoto, no ProRAW, no ProRes. It isn't Pro, it's just Duo. PetaPixel called the main camera "the epitome of fine."
Then 14 reviewers hit the same wall: the apps aren't ready. I'm sure they will be eventually, who knows when. Until then it's shrunk iPhone apps floating in dead space, like buying a widescreen TV and watching everything through a porthole.
It's 254 grams, basically a hamster living in your pocket, and Gizmodo said the only solution was to wear a belt again. There's no Face ID, and 12 reviewers missed it. Touch ID on the side button has a deliberate delay, and in winter it fails with gloves.
Apple moved the dock and the status bar to a rail down the right edge, and 6 reviewers said some version of "you have to relearn the iPhone." Well, if you've got to relearn the iPhone, why not learn Android?
Then there's the price, which 12 reviewers brought up. HardwareZone did the math: for that money, you could get an iPhone 18 Pro and an iPad mini with change to spare, the two devices this phone is trying to replace. And if you crack the inside screen without AppleCare, that's $999.
You can see the crease at an angle, less than Samsung but not zero, and I don't know why people make it a big deal, because it's a foldable phone and it's going to have a crease.
The Verge said expensive, full of compromises, and switching back to a normal phone feels like a relief. MKBHD says most people shouldn't buy it. That's $2,000, more than a month's rent for most Americans as CNN pointed out, for a phone whose own reviewers keep using the same word: compromise.
If you're going to get an iPhone Duo, I only have one question for you. Why?
15
0
10
Yes, the LockerGnome tech news website has returned (with a modern twist)! "Tech news and nostalgia for today's technology professionals."
4
3
33
One hobbyist built a working laptop around a 6502, the chip inside the Apple II...
A maker who goes by TechPaula built a laptop around a 6502, the 8-bit chip from the 1970s that Hackaday points out was in the Apple II, and now they've finished a second one. They wrote on their ch ramblings, on August 30 that the LT6502b is done, in a translucent orange case. It runs at 10MHz, comes with EhBASIC and a monitor program, and stores its files on a CompactFlash card.
Hackaday covered the first version in February: a 9-inch foldable screen, a 10,000 mAh battery, 46 KB of RAM, a mostly 3D-printed case and a custom keyboard with a little 8-character LED display built in. That one was chunky. The project page on Codeberg says the new case is 31mm tall, against 70mm for the old one, and TechPaula's blog says it's half the thickness.
The second one is nicer to live with. The screen grew to 10.1 inches at 1024 by 600, up from 9 inches at 800 by 480, and it's IPS now. They added a SID emulator for sound and a WiFi modem, and the Codeberg notes say they've connected to a few BBSes with it. (A 6502 dialing a BBS over WiFi in 2026 is the best kind of anachronism.) They dropped the expansion slot to keep it sleek.
My favorite part of the post is a list of things TechPaula would change. The WiFi modem is too slow and a bit flaky for real use, the speaker is a bit anemic and the heat insert holes are very weak. If you've ever wanted to build a computer from the chip up, I'd read that list before you order a single part, since it's the stuff nobody puts in the glamour shots. The code and PCB files are on Codeberg; the old GitHub page now says the project moved there.
It's orange. I want one.
A maker who goes by TechPaula built a laptop around a 6502, the 8-bit chip from the 1970s that Hackaday points out was in the Apple II, and now they've finished a second one. They wrote on their ch ramblings, on August 30 that the LT6502b is done, in a translucent orange case. It runs at 10MHz, comes with EhBASIC and a monitor program, and stores its files on a CompactFlash card.
Hackaday covered the first version in February: a 9-inch foldable screen, a 10,000 mAh battery, 46 KB of RAM, a mostly 3D-printed case and a custom keyboard with a little 8-character LED display built in. That one was chunky. The project page on Codeberg says the new case is 31mm tall, against 70mm for the old one, and TechPaula's blog says it's half the thickness.
The second one is nicer to live with. The screen grew to 10.1 inches at 1024 by 600, up from 9 inches at 800 by 480, and it's IPS now. They added a SID emulator for sound and a WiFi modem, and the Codeberg notes say they've connected to a few BBSes with it. (A 6502 dialing a BBS over WiFi in 2026 is the best kind of anachronism.) They dropped the expansion slot to keep it sleek.
My favorite part of the post is a list of things TechPaula would change. The WiFi modem is too slow and a bit flaky for real use, the speaker is a bit anemic and the heat insert holes are very weak. If you've ever wanted to build a computer from the chip up, I'd read that list before you order a single part, since it's the stuff nobody puts in the glamour shots. The code and PCB files are on Codeberg; the old GitHub page now says the project moved there.
It's orange. I want one.
1
1
1
On this day in 1969, the internet's first message crashed after two letters. UCLA to SRI, 10:30 at night: Charley Kline typed 'lo' of 'login' and the system died. Lo, and behold. (Sources: ICANN, University of California)
1
1
5
Five years ago today I unboxed the Pixel 6 Pro, six months after I'd last used a Pixel as my go-to phone. youtu.be/et9sH97TnYs
1
0
2
Today is National Cat Day - 3 Facts!
• Founded in 2005 by Colleen Paige, and it lands every October 29.
• About 3 million cats entered U.S. shelters and rescues in 2025.
• About 2.2 million cats were adopted from them that year.
(Sources: National Cat Day, ASPCA)
• Founded in 2005 by Colleen Paige, and it lands every October 29.
• About 3 million cats entered U.S. shelters and rescues in 2025.
• About 2.2 million cats were adopted from them that year.
(Sources: National Cat Day, ASPCA)
1
2
5
Battery life versus screen real estate. Personally, I would trade battery life for screen real estate, even with an Android device.
0
0
1
iPhone Duo sold out until January? Don't pay a scalper...
Apple's first foldable starts at $1,999, and Gizmodo reported that its ship dates slipped from launch day, October 23, to mid-November in under five minutes. By 5:30 a.m. Pacific, MacRumors saw every configuration showing ed that first wave, you're looking at three months, a store pickup or a reseller.
The resellers were ready before you were. Gizmodo found one X user selling an "auto checkout" bot to paying members, and bragging about "hundreds of dollars in potential profit per unit" without getting out of bed. That profit is the markup somebody pays because they didn't want to wait, and I wouldn't fund the bot that beat you to the button.
January is an estimate, and MacRumors says as much: nobody knows how many phones Apple had on day one, and the dates reflect supply and demand together. Gizmodo cites a Securities Times report that Apple raised production 30%, to an estimated 10 million units for 2026. iDrop News says Apple Stores typically get shipments at least twice a week.
So if you've got a pre-order in, keep it, and watch in-store pickup in the Apple Store app instead of refreshing the ship date. Apple told people to check back for store availability on Saturday, October 24, the day after launch. MacRumors expects extremely tight inventory for walk-in customers, so I'd check that app every morning for the first couple of weeks. And if somebody is selling a Duo for well over $1,999, they're selling you a spot in the same line Apple runs for free.
I already made a whole YouTube video about why I'm not buying one. If you are, January costs you time and nothing else.
Apple's first foldable starts at $1,999, and Gizmodo reported that its ship dates slipped from launch day, October 23, to mid-November in under five minutes. By 5:30 a.m. Pacific, MacRumors saw every configuration showing ed that first wave, you're looking at three months, a store pickup or a reseller.
The resellers were ready before you were. Gizmodo found one X user selling an "auto checkout" bot to paying members, and bragging about "hundreds of dollars in potential profit per unit" without getting out of bed. That profit is the markup somebody pays because they didn't want to wait, and I wouldn't fund the bot that beat you to the button.
January is an estimate, and MacRumors says as much: nobody knows how many phones Apple had on day one, and the dates reflect supply and demand together. Gizmodo cites a Securities Times report that Apple raised production 30%, to an estimated 10 million units for 2026. iDrop News says Apple Stores typically get shipments at least twice a week.
So if you've got a pre-order in, keep it, and watch in-store pickup in the Apple Store app instead of refreshing the ship date. Apple told people to check back for store availability on Saturday, October 24, the day after launch. MacRumors expects extremely tight inventory for walk-in customers, so I'd check that app every morning for the first couple of weeks. And if somebody is selling a Duo for well over $1,999, they're selling you a spot in the same line Apple runs for free.
I already made a whole YouTube video about why I'm not buying one. If you are, January costs you time and nothing else.
4
0
5
Windows 98 just got USB 3 and Thunderbolt drivers, 28 years late...
Windows 98 went on sale on June 25, 1998, according to Wikipedia, and in 2026 it can read a Thunderbolt NVMe drive. Yeo Kheng Meng released a driver called xHCI98 that gives Windows 98 SE, ME, 2000, XP, Vista th experimental support for USB4 and Thunderbolt 3 and 4. Hackaday headlined it "Thunderbolt Strikes Windows98," which is about right.
It started with a laptop he didn't want to give up. Hackaday reported in September that he likes Windows 98, but he also likes his 2020 ThinkPad, which only has an xHCI USB controller, and nothing existed to make the two get along. Hackaday says he decided to do it himself, or at least guide LLMs through writing the drivers. On his blog, Yeo writes that this time he decided to go further and give the driver both USB 3 and USB4/Thunderbolt support.
The speeds are the fun part. The project's GitHub page reports about 210 MB/s read and 256 MB/s write on Windows 98 SE, against 1,151 MB/s read and 1,013 MB/s write on Windows 7, and says Windows 98's own storage stack is the bottleneck, not the Thunderbolt link. Hackaday adds that Windows 98's 64 KB maximum request limit likely holds USB 3 to around 200 MB/s.
If you've got a retro PC project, I'd read the warnings before the speed chart. Yeo writes that apart from the Thunderbolt 3 traffic he captured on his ThinkPad P14s Gen 1, everything else is still theoretical, and the USB4 side has only been tested against an emulator. Hackaday says the Thunderbolt support isn't as stable as the USB 3 part. The GitHub page tells XP users to keep a PS/2 or built-in keyboard handy, and warns that unplugging a drive can lose cached writes. I'd try it on a machine you can afford to reinstall, with a copy of anything you care about somewhere else.
Windows 98 never had it this good.
Windows 98 went on sale on June 25, 1998, according to Wikipedia, and in 2026 it can read a Thunderbolt NVMe drive. Yeo Kheng Meng released a driver called xHCI98 that gives Windows 98 SE, ME, 2000, XP, Vista th experimental support for USB4 and Thunderbolt 3 and 4. Hackaday headlined it "Thunderbolt Strikes Windows98," which is about right.
It started with a laptop he didn't want to give up. Hackaday reported in September that he likes Windows 98, but he also likes his 2020 ThinkPad, which only has an xHCI USB controller, and nothing existed to make the two get along. Hackaday says he decided to do it himself, or at least guide LLMs through writing the drivers. On his blog, Yeo writes that this time he decided to go further and give the driver both USB 3 and USB4/Thunderbolt support.
The speeds are the fun part. The project's GitHub page reports about 210 MB/s read and 256 MB/s write on Windows 98 SE, against 1,151 MB/s read and 1,013 MB/s write on Windows 7, and says Windows 98's own storage stack is the bottleneck, not the Thunderbolt link. Hackaday adds that Windows 98's 64 KB maximum request limit likely holds USB 3 to around 200 MB/s.
If you've got a retro PC project, I'd read the warnings before the speed chart. Yeo writes that apart from the Thunderbolt 3 traffic he captured on his ThinkPad P14s Gen 1, everything else is still theoretical, and the USB4 side has only been tested against an emulator. Hackaday says the Thunderbolt support isn't as stable as the USB 3 part. The GitHub page tells XP users to keep a PS/2 or built-in keyboard handy, and warns that unplugging a drive can lose cached writes. I'd try it on a machine you can afford to reinstall, with a copy of anything you care about somewhere else.
Windows 98 never had it this good.
0
0
1
The man who put Lemmings on the Palm Pilot just put his whole Palm OS toolbox on GitHub...
Aaron Ardiri, the developer behind the Palm OS port of Lemmings, just published about a gigabyte of old Palm OS development material on GitHub. The Register reported that the files had sat for more than 25 years. What got him to open the vault was a pull request.
Somebody had sent a change to his old Lemmings source code, adding HiRes+ display support for the Palm Tungsten T3. Ardiri told The Register he had no idea people were still improving that software more than two decades later. (Every Palm owner I knew had Lemmings on it, so this tracks.) Then he found the Palm fans on Reddit and Discord hunting for exactly the kind of stuff sitting on his backup drives. He also told The Register he recently handed more than 50 old Palm devices to Dmitry Grinberg, who works on keeping Palm OS alive.
The archive's GitHub page lists SDKs, developer conference presentations, utilities, programming examples and a limited release of Palm OS source code for versions 3.5, 4.0 and 5.0. He left out the things he didn't have permission to hand around, like ROM images, firmware, simulators and commercial apps, and The Register notes that being on GitHub doesn't make any of it open source. The Lemmings code itself has been out for a while; PalmDB lists the full open-source package, port by Ardiri and Chip Kerchner, submitted back in November 2021, and it carries no affiliation with Psygnosis.
If you've still got a Tungsten or a Zire in a drawer, I'd dig it out. Fans are still patching Lemmings for the Tungsten T3, the full package is free on PalmDB, and if you want to write something of your own, the archive hands you the original SDKs.
Grab your stylus. It's been a while.
Aaron Ardiri, the developer behind the Palm OS port of Lemmings, just published about a gigabyte of old Palm OS development material on GitHub. The Register reported that the files had sat for more than 25 years. What got him to open the vault was a pull request.
Somebody had sent a change to his old Lemmings source code, adding HiRes+ display support for the Palm Tungsten T3. Ardiri told The Register he had no idea people were still improving that software more than two decades later. (Every Palm owner I knew had Lemmings on it, so this tracks.) Then he found the Palm fans on Reddit and Discord hunting for exactly the kind of stuff sitting on his backup drives. He also told The Register he recently handed more than 50 old Palm devices to Dmitry Grinberg, who works on keeping Palm OS alive.
The archive's GitHub page lists SDKs, developer conference presentations, utilities, programming examples and a limited release of Palm OS source code for versions 3.5, 4.0 and 5.0. He left out the things he didn't have permission to hand around, like ROM images, firmware, simulators and commercial apps, and The Register notes that being on GitHub doesn't make any of it open source. The Lemmings code itself has been out for a while; PalmDB lists the full open-source package, port by Ardiri and Chip Kerchner, submitted back in November 2021, and it carries no affiliation with Psygnosis.
If you've still got a Tungsten or a Zire in a drawer, I'd dig it out. Fans are still patching Lemmings for the Tungsten T3, the full package is free on PalmDB, and if you want to write something of your own, the archive hands you the original SDKs.
Grab your stylus. It's been a while.
2
1
2
Don't sleep on AI's ability, and I use that term in an umbrella capacity, these tools' ability to help lift the cognitive load.
0
0
0
Vibe Coding & AI News to Use: October 28, 2026…
Dedica launches an agent-readiness website checker on Product Hunt.
getdedica.com
@BotGauge brings agent red-teaming, replay and regression checks to Product Hunt.
botgauge.com
@ParthJadhav8 offers a local screcoding agents.
parthjadhav.com/products/app-s…
@seshch launches its dependency-license compatibility CLI on Product Hunt.
github.com/shcherbak-ai/l…
Our Silly Games offers browser-game hosting for games built with coding agents.
oursilly.games
Yardlet launches a Mac workspace for existing coding-agent CLIs.
yardlet.app
Mudroom previews sandboxed coding-agent sessions with diff review and undo.
github.com/Kernel-Hunter/…
Trigger.dev showcases webhooks for AI agents on Product Hunt. @triggerdotdev
trigger.dev
Google introduces Gemini 4 Carbon.
blog.google/innovation-and…
NVIDIA introduces promptable multilingual Nemotron 3.5 Transcribe and Granary v2.
huggingface.co/blog/nvidia/ne…
Try this image style:
Take the attached photo and redraw it as Flung-paint action canvas: eyes two clean bone-white ovals with heavy black paint-drop pupils behind rectangular glasses drawn as two bold splashed frames, nose one confident swipe of a loaded brush, mouth a firm closed smile in a single thick arc, hair black spikes flung upward off the scalp like paint thrown from a spinning brush, and hands and body one solid silhouette with enamel drips flying off the shoulders, both hands mid-motion, each wholly in frame. Palette: cadmium red, bone white, ink black, gym-floor tan. Full-figure wide shot inside a boxing ring: Chris at left mid-footwork landing a light playful jab, the friendly boxy robot training partner at right with padded arms up in guard and a happy gauge on its chest, thick flung paint arcs tracing both fighters' movement, one ring post and the ropes framing the far edges, plain off-white ground with a few paint splatters, his face toward the viewer in three-quarter view. Keep the person recognizable. Square, no text.
Meet other builders at ctrl+alt+create live Builder Night and Builder Day events.
luma.com/ctrlaltcreatel…
ctrlaltcreate.live
Dedica launches an agent-readiness website checker on Product Hunt.
getdedica.com
@BotGauge brings agent red-teaming, replay and regression checks to Product Hunt.
botgauge.com
@ParthJadhav8 offers a local screcoding agents.
parthjadhav.com/products/app-s…
@seshch launches its dependency-license compatibility CLI on Product Hunt.
github.com/shcherbak-ai/l…
Our Silly Games offers browser-game hosting for games built with coding agents.
oursilly.games
Yardlet launches a Mac workspace for existing coding-agent CLIs.
yardlet.app
Mudroom previews sandboxed coding-agent sessions with diff review and undo.
github.com/Kernel-Hunter/…
Trigger.dev showcases webhooks for AI agents on Product Hunt. @triggerdotdev
trigger.dev
Google introduces Gemini 4 Carbon.
blog.google/innovation-and…
NVIDIA introduces promptable multilingual Nemotron 3.5 Transcribe and Granary v2.
huggingface.co/blog/nvidia/ne…
Try this image style:
Take the attached photo and redraw it as Flung-paint action canvas: eyes two clean bone-white ovals with heavy black paint-drop pupils behind rectangular glasses drawn as two bold splashed frames, nose one confident swipe of a loaded brush, mouth a firm closed smile in a single thick arc, hair black spikes flung upward off the scalp like paint thrown from a spinning brush, and hands and body one solid silhouette with enamel drips flying off the shoulders, both hands mid-motion, each wholly in frame. Palette: cadmium red, bone white, ink black, gym-floor tan. Full-figure wide shot inside a boxing ring: Chris at left mid-footwork landing a light playful jab, the friendly boxy robot training partner at right with padded arms up in guard and a happy gauge on its chest, thick flung paint arcs tracing both fighters' movement, one ring post and the ropes framing the far edges, plain off-white ground with a few paint splatters, his face toward the viewer in three-quarter view. Keep the person recognizable. Square, no text.
Meet other builders at ctrl+alt+create live Builder Night and Builder Day events.
luma.com/ctrlaltcreatel…
ctrlaltcreate.live
3
1
5
Somebody got Android apps running on a Mac with no emulator, and they show up in the Dock...
A developer who goes by kksimp put a project on GitHub, Mac dot apk, that runs Android apps on Apple Silicon Macs with no emulator and no virtual machine. The project's page says the macOS processes, get their own windows with a back button next to the traffic lights, and sit in your Dock like any other app. MacGeneration tried it and saw the same Dock icons and standard windows.
It works because your Mac and an Android phone already speak the same language. El Solitario explains that both chips share the ARM64 architecture, so there's nothing to translate, and a recompiler converts the code of older ARM32 apps to ARM64 before running it. The project says apps draw through the Mac's GPU and run on the CPU at native speed. You install one by double-clicking the APK file or dragging it onto the app's window, which is about as much ceremony as a Mac app from a disk image.
It's an alpha, and the numbers say so. The project lists 65 apps tested: 51 fully playable, 10 partly working and four that don't run yet. MacGeneration had TikTok crash right at startup, while the game Mindustry ran fine apart from a few sound bugs. Appsystem flags games with anti-cheat systems as likely trouble. You'll need macOS 26.4 Tahoe or newer, and Intel Macs are out.
The part I'd read twice is the security note. The project's own page says there's no phone-style sandbox around these apps: an Android app runs as native code with your Mac account's permissions, and it can read and write your files the way any Mac app can. So if you're curious (and I am), I'd stick to APKs from a source you already trust, and only the ones you'd let near your Documents folder.
It's fun to try, as long as you treat every APK like a Mac app somebody handed you on a thumb drive in a parking lot.
A developer who goes by kksimp put a project on GitHub, Mac dot apk, that runs Android apps on Apple Silicon Macs with no emulator and no virtual machine. The project's page says the macOS processes, get their own windows with a back button next to the traffic lights, and sit in your Dock like any other app. MacGeneration tried it and saw the same Dock icons and standard windows.
It works because your Mac and an Android phone already speak the same language. El Solitario explains that both chips share the ARM64 architecture, so there's nothing to translate, and a recompiler converts the code of older ARM32 apps to ARM64 before running it. The project says apps draw through the Mac's GPU and run on the CPU at native speed. You install one by double-clicking the APK file or dragging it onto the app's window, which is about as much ceremony as a Mac app from a disk image.
It's an alpha, and the numbers say so. The project lists 65 apps tested: 51 fully playable, 10 partly working and four that don't run yet. MacGeneration had TikTok crash right at startup, while the game Mindustry ran fine apart from a few sound bugs. Appsystem flags games with anti-cheat systems as likely trouble. You'll need macOS 26.4 Tahoe or newer, and Intel Macs are out.
The part I'd read twice is the security note. The project's own page says there's no phone-style sandbox around these apps: an Android app runs as native code with your Mac account's permissions, and it can read and write your files the way any Mac app can. So if you're curious (and I am), I'd stick to APKs from a source you already trust, and only the ones you'd let near your Documents folder.
It's fun to try, as long as you treat every APK like a Mac app somebody handed you on a thumb drive in a parking lot.
1
0
1
On this day in 2009, a gamer called my live show to say Windows 7 finally fixed his games. I am still taking calls about the next Windows. youtu.be/TEcHmqA62x8
0
0
3