A new @HaskelInterlude episode is out, this time with @kmett!
We talked about design principles for Haskell libraries, category theory, cache-oblivious algorithms and his recent work involving Haskell, C++, and FPGAs.
Tune in!
https://t.co/KYXbv8SO8e
I've been building so much with Spacetime over the past few months that it's time I showed some details. Here's a TypeScript build in which I stumble on Wikipedia's edit log and turn it into a real-time, pubsub, data-streaming leaderboard.
https://t.co/3ZGDrqwNf5
I'm super excited that @krisajenkins has joined us to help spread the good word of Spacetime!
Check out Kris' first video on our channel! He's coming out of the gate strong.
https://t.co/e0JA18KAc6
@starwardshar@Ronanchamberss This is the usual British story. The NHS has competent dull projects and people who want to see them through. The Ministers want to be SV execs.
I stand by that some of the best ideas in Computer Science can be found from failed research of the late 90s.
DataScalar is no longer around, but they would have made a killing today.
They created an architecture which was, to put it simply, a memory centric execution model. You trade excess threads for reduction of data movement.
But *unlike* multithreading the way we think about it today, you’d have *redundant*, sequential programs running multiple times.
At the time, DataScalar was making the (risky!) bet that future processors would have arithmetic become virtually free compared to data movement. You deliberately waste (tons) of CPU cycles, on an otherwise unaware application, where the CPU that “wins” is whichever one happens to be closer to the memory that’s needed.
Instead of designing a better prefetcher, you just race all the threads against each other, the winner (local memory owner) one-way broadcasts, and the other cores get the result without making a conventional remote request.
Today, you’d call it…an odd form of near-data processing. What I find really interesting is in the present era most redundant processing is done in a lockstep / safety sense (aerospace, etc), not a performance / latency sensitive sense.
Weirdly, the technique isn’t that far off from the data racing I did in my TailSlayer project to reduce RAM refresh latency stalls.
I still find it surprising how few people in the US talk about penpot. When you visit Europe, everyone is using it.
It is open source: https://t.co/57eG5uRulB
@AndrewOrlowski The FDP has a whole mush of other stuff in it, but it all seems to be copies of things that exist elsewhere. No one really uses it much because it means double keying data.
@tom_watson I’ll stick by my view that Palantir are slightly mediocre software with good but pricey consultancy who deliberately politicise themselves for marketing purposes.
https://t.co/q4IXbcLgBR
I left high frequency trading to join @fuseenergy and had to sit out SIX MONTHS of gardening leave (doing no work).
I believed in Fuse and was desperate to witness its early days (+ I’m not built for prolonged holidays 😅).
I now lead engineering here. I believe in the right to EXIT. If this isn’t the best place in the world for my engineers, they should be free to leave.
I'm at OpenAI's DevDay event in San Francisco today - as I have for the past three DevDay events, I'm running a live blog where I'll be posting updates during the keynote, which starts in five minutes https://t.co/X9DLSS0xbR
2/ Britain is unusually permissive in its treatment of non-competes, notice periods, and garden leave. These clauses reduce competition, undermine innovation, and lower wages.
They are particularly damaging in the AI sector, where a month locked out of the labour market leaves you far behind.
https://t.co/BjwSZplo5v