Haha. Thanks man. I dimly remember it took some messing with the arm assembly to get the scrolling fast enough to make the v-sync. And also that we used the hardware mouse pointer sprite to draw the player's spaceship, to get an extra couple of colours. Good times!
It was something entirely new, of course. In fact it wasn't called a BBC Micro - however it was "endorsed" by the BBC - had the owl logo on it etc. Was indeed a cool machine, and streets ahead of anything else at the time. Typical UK tech story...
I always loved Acorn computers. My schoolfriend and I released a commercial game on the Archimedes, and in 1994 I wrote a 3D demo suite for Acorn's new RiscPC machine (powered by ARM, of course). The good old days of hacking around!
The Archimedes was called a "BBC Micro" because it was part of the BBC's home computing initiative, but architecturally, it had nothing to do with the original BBC Micro.
I agree. I think when something is at the cutting edge, and you're creating effects no-one has ever seen before - (the first Jurassic Park, say) - people really go the extra mile to make it look as great as possible. Fast forward to today, when CG is pretty much a commodity - and it's often a lot more about getting it done on time and within budget.
We have fingers and a keyboard / mouse, because that's the best thing we've come up with. If we could output binary to a bunch of program interfaces directly (perhaps via some neuralink type thing) we would surely take that option.
Yes, perhaps that's right. Seems pretty dishonest to me though. He's a "serious Harvard scientist", so people listen when he tells us that some rock they've spotted may be an alien spaceship. Even though there is surely zero evidence for such an assertion. What the hell is he doing?
An issue I have with this, is that it poses two rather contrary hypotheses.
1. When considering a career, people do a bad job of unpacking the detailed day-to-day activities that make it up.
2. When shown these detailed day-to-day activities people can do an excellent job of assessing whether they would be "happy" doing it or not.
Why would a person who is so bad at making the leap from the vague notion of being a professor to imagining the real-world actualities of that job, suddenly gain 100% detailed insight into whether or not those actual tasks would make them happy?
The lack of photorealistic fidelity gives your brain a bit of room to use imagination to fill in the blanks in your internal model. This fosters a certain type of engagement with the content that you don't get with photorealistic images.
The model starts every conversation as a blank slate, so providing a thorough context regarding the problem you want it to solve seems a fairly obvious preparatory step tbh. How else is it supposed to know what to do? I agree that "prompt" is probably not quite the right word to describe what is necessary though - it feels a bit minimal and brief. "Context engineering" seems a bit overblown, but this is tech. and we do a love a grand title.
I'm a software engineer by trade, but have played bluesy / rock guitar since I heard Jimi Hendrix as a teenager. I try to run a band, to fuel my mid-life crisis but I've come to the conclusion that it's essentially a hobby, not a device for making money.
There's a certain change of perspective with modern AI (by "modern" I mean Resnet and beyond). When I was deep into neural nets in the 1990s, they weren't that large, and I would think of them in terms of the number of weights and nodes - but modern deep learning seems to have has moved up a few levels of abstraction. (I stepped away from the field for a while). And there's a certain understanding people seem to have now regarding the "gradient flow" through the net and why certain architectures work well (Resnet, Unets etc). I must say I'm finding it tricky to shift into this new level of thinking. Also Transformers - still looking for an intuitive sense of how they work, haha.
Tech people like it because it isn't good enough to completely replace them yet. The sophisticated, coherent architecture of a well designed system is (for now) still beyond the LLMs, so for tech people, it's still just a wonderful tool. But give it another year, and the worm will turn.
Back in the early 90s I did a version of Bresenham's algorithm that would rasterize the hyperbolic curves that perspective-correct texture mapping required. It worked correctly though the technique of just doing a division every n pixels and linearly interpolating won out in the end, if I recall.
I've never been quite clear on how Splats encode specular (directional) effects. Are they made to only be visible from a narrow field of view (so you see a different splat for different view angles?) or do they encode the specular stuff internally somehow?
It's interesting - had they acted this schadenfreude out at all? Or did this confession come completely out of the blue. From what you said, it is the latter - and if so, should this person be judged for this unappealing inner thought, if they take pains not to let that thought influence their behaviour or actions?