
Local-AI enthusiast. Built things way before ES6 and Docker-era. Tweets are NFA. GU #1884 🦇🔊
Mainnet
Joined August 2021
- Tweets1K
- Following277
- Followers896
- Likes7.6K
The other side is you can build your pet projects ultra fast, even the crazy ideas are finally getting some attention. Use two sides of a coin, not the only one.
I am done with this shit. It is over. The state of engineering right now is horrible. It has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, reports, etc., everything is made by Claude Code. Nobody on my team likes this. They are being forced to ship as much as they can. I have heard multiple times from higher management that pushing code is not a bottleneck, so why are we slow? People are working 12 to 13 hours a day just to press enter. Nobody is reading anything. Humans in corporate are doing nothing on their own. Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude. There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore. Everything is done by LLMs. It is so soul-sucking. I would not mind it, to be honest, if we were at least given the time to check out the code and see what is going where. But no, the goal is to just ship. No matter what happens.
63
Prices for 5090 are out of control. I was thinking that I was heavily overpaying during 2025, but turns out it was an investment 😂
39
xdm99 retweeted
my favorite AI slop must be coding agents adding backward compatibility and legacy fallbacks for an app that isn't even deployed yet
xdm99 retweeted
Replying to @firesidealpha
Sudden Clarity Sam realizing that real-world dynamics equations have inertia and friction terms.
(Particularly when they concern large systems like human societies).
If only the singularitarians, AI doomers, AGI-is-near crowd, and other SV cultists could realize that....
exactly, the bare minimum of being independent in the current agentic era is 128gb unified ram or equivalent for gpu+ram offloading (approx. 2x3090 + 96gb ddr5 @ 6000+)
All you need is a single DGX Spark, or a 128GB RAM Mac.
You can run uncensored DeepSeek V4 Flash with it.
This brings you actual freedom.
Freedom from censorship, control, and weekly usage limits.
No more begging for weekly limit resets, no more worrying about service outages, and no more getting your account banned for asking a stupid question like "what is mitochondria".
Own your intelligence on your desk.
3
244
Do you know which comes next after Skills in markdown? Entire mindset in vector format. And a builder to create it. With a grain of negative perks to hit better randomness. Like in Fallout 2 character creation screen.
4
139
xdm99 retweeted
Unlimited mathematical proofs but limited skilled mathematicians who can verify them
Unlimited code but limited skilled programmers to review them
Unlimited videos but limited attention to watch them
Unlimited lawsuits but limited courts/judges/lawyers to adjudicate them
When the creation of something goes to zero, the cost moves elsewhere…
Not all the Chinese providers have the same vibe. DeepSeek and MoonshotAI under-market and over-deliver consistently. Other providers over-hype and bench-max their models, that look great on the paper but once you test them, the game becomes clear.
xdm99 retweeted
I can't believe this works, but I got DeepSeek-V4-Flash (284B params) running on a Raspberry Pi 5 (8GB edition) at >1tok/s @ ~8W during full-tilt inference! It uses an untouched copy of @antirez's GGUF. Took 160+ experiments over 5 days between GPT-5.5 xhigh and Opus 4.8 max.
Hey @grok start playing The Doors - The End
❗️🚨 BREAKING: Researchers used Mythos Preview to find the first public macOS kernel memory corruption exploit on Apple's M5 silicon, they give a glimpse into Mythos say it’s really powerful.
Apple spent five years and an estimated several billion dollars building Memory Integrity Enforcement (MIE), the hardware-assisted memory safety system built around ARM's MTE. It was the flagship security feature of the M5 and A19, designed specifically to kill the entire memory corruption bug class.
Researchers from Calif built a working exploit in five days.
According to Apple's own research, MIE disrupts every public exploit chain against modern iOS, including the recently leaked Coruna and Darksword kits. Calif walked into Apple Park this week and handed over the report in person.
Full 55-page technical report drops after Apple patches the vulnerability.
2
1
102
This is pure gold. A perfect example of why this local-first AI scene is so amazing. Contributors gonna contribute. Generous people gonna donate. Local LLMs go brrrrrr. Everyone is happy.
Announcing with gratitude that @audreyt just gifted me an M5 Max 128GB MacBook Pro! It will let me develop DwarfStar4 (DS4) further on newer Metal hardware, and experiment with distributed inference splitting Q4 quants across the M3 Max and the M5 Max. Thanks so much, Audrey!
1
2
84
xdm99 retweeted
We used to go to a special website, ask strangers for help with programming, and get humiliated in return
If you bought 256GB DDR5 in the first half of 2025 just to have some offloading space for DeepSeek and Kimi models, you basically won 2025, 2026, 2027 and probably the rest up to 2030.
Coding plans are subject to change. Ram sticks are the real solid asset. Choose wisely.
If you bought GLM annual coding plan during the December promo, you basically won 2026. 😅
You'll stay on the legacy plan (with no 5h and weekly limits).
That was an amazing deal!
1
1
183
























