Benathon retweeted
"I worked HARD for the fans, every day."
I now have permission to share something I hoped would be revealed with much more fanfare.
I spent more than 3 years as one of the writers of #MassEffect4
It was one of the greatest privileges & GLORIOUS challenges of my life. I smiled every day. I worked HARD for the fans, every day.
So when I say I am devastated by the end of #MassEffect I REALLY mean it. I don’t know how to express how gutted I truly am.
Benathon retweeted
I just had AI write an entire RISC-V core, put it on an FPGA, compile C code into a custom ELF format, boot the core, and run the program.
I read ZERO of the code. Someone come up with a good name for the core.
Benathon retweeted
As I cross over the 200k of C++ in my game, I observe the following trend for how much I use AI to write the code for me between C++, shaders and Blender add-ons:
C++ - 99% hand-written, 1% AI
shaders - 50% - 50%
Blender add-ons - 1% hand-written, 99% AI
Benathon retweeted
Photographer Brendan Larsen left his camera recording the night sky over Mt Taranaki, New Zealand, for nearly two hours, capturing 578 photographs.
The result? An estimated 14,450 satellite trail appearances, representing approximately 1,400 individual satellites.
Benathon retweeted
Unlocking everything on the AMD CPU with DRAM scrambling (@xoreaxeaxeax)
github.com/xoreaxeaxeax/skit…
#infosec
Benathon retweeted
Bypassed the native wireless stack on ESP32-S3 & S31 to achieve arbitrary IQ transmission! Effectively turning these cheap IoT chips into SDR transmitters for custom RF protocols. 🤯Cleaning up the code with some example on GitHub today! #ESP32S3 #S31 #SDR #RF #HardwareHacking
Benathon retweeted
This board is no longer available for sale - 12 WiFi radios and 24 antennas - reduced down to 4 external antennas.
build/principle of antenna ->LNA->RF Combiner->6 antenna wires for 3 radios doing it 4 times for the reduced attention of having a lot of antennas on the vehicle.
Benathon retweeted
Me 5 minutes after buying a Tesla Roadster xitter.cf/ssssscca/status/211098…
Benathon retweeted
people are starting to realize.
there *are* no moats left. the only true moat in existence was intelligence, and now that's out of the equation.
Benathon retweeted
For the cost of 8 years of $200 Claude subs, you can force your team to use shitty quantized versions of DeepSeek v4 Flash running at 30tps. Incredible deal!
Local AI is good enough for a startup. A 10-person startup built its own 128GB-VRAM AI server for $18,000.
One Threadripper box sitting locally with
🎮 4× Radeon AI PRO R9700 32GB
💾 128GB total VRAM
🧠 Qwen3.8 + DeepSeek V4 Flash
👥 16 concurrent AI sessions
And they're actually pushing it.
Qwen3.8-Flash-Next with 16 simultaneous sessions:
📚 48K context/user
⚡ 538.5 tps aggregate decode
At 128K context per user
⚡ 365 tps aggregate decode
The GPUs are even undervolted to under 210W each.
Link in ALT
ALT https://www.reddit.com/r/LocalLLaMA/comments/1x2u742/building_a_4x_r9700_setup_for_a_10_person_startup/