Techy building the future. Entrepreneur, SWE/PA🤓 e/acc, SpaceX, and Tesla 🚀 Real Estate Investor. Cybertruck supervisor 📐
Joined September 2010
- Tweets11.1K
- Following2K
- Followers1.3K
- Likes32.4K
The current bar to surpass is the distracted meatsack with ears/eyes and vibration sensor that sits in 99.999% of driver's seats.
I get that at some point, to reach beyond that, we may reach for active sensors, but one step at a time.
General Motors says it will not adopt a vision-only approach to autonomous driving, saying that dropping radar is a "weird religious stance that ignores physics."
"If you scale back to vision-only, physics suggests you’ll be in trouble in certain situations. Radar is so cheap that dropping it doesn’t make sense—dropping it is a weird religious stance that ignores physics," GM Chief Product Officer Sterling Anderson said.
GM says it plans to launch an autonomous driving feature for the highway by 2028.
Full article: insideevs.com/news/813467/gm…
For folks wanting @bot integration with iMessage, that data is stored in a local SQLite DB, so you can just ask your bot for a monitor routine... or iterations thereof.
Starting to use local models for activities I don’t want to ‘waste’ foundation model tokens on: Important email checker/classifier. Security cam, ‘is it real’ alert debouncer, etc…
False positives get added to a list that I have a routine to tweak.. somewhat self-improvement.
D3-flash as a decider and qwen3.8 as a vlm. Hopefully omni gets a lms-compat quant soon
When an AI can do a decent job of naming a product, I'll call it AGI. (it's sometimes still off-by-one so...)
This feels like the X number to me. Yeah, I get at there's probably a niche there for folks who keep their follow count low (??), but for the majority of users they just use the app? (and follow to chat?)
By popular request, you can now text your dot.
Give it a task to take off your plate or something to keep an eye on, using text, photos, or even voice notes. It will text you proactively with relevant updates.
Safe, secure texting via iMessage or RCS. You’re the only one who can text your dot, and your dot can only text you. Rolling out now as a preview for Pro users in the US.
I find it funny that one can't pay the @GooglePlay Developer registration fee with @GooglePay ... 🤷
Yeah. If there could be four changes to @bot I’d want the task list, auto-creation of task scripts when possible, ability to fire and forget then pickup tasks to/from different processes, and token-less auto mgmt of the bot box, or to indicate to use a user’s box completely. Gimmie a bot container if needed.
Maybe I'm wrong on #3 and it already creates a monitor and comes back to it but it seems like I'm burning tokens on backgrounded tasks.
Claude says things like "The monitor will wake me when they finish," but I haven't seen that with bot.
Sean McLellan retweeted
1/4
Meet Qwenputer: Qwen3.5-4B with a full RISC-V CPU inside its weights, using the model's own Gated DeltaNet layers as RAM.
It runs C and MicroPython in its forward passes, no external runtime.
On DeepMind math: base 55–90%, base + Python tool 95–100%, Qwenputer 97.5–100%.
Sean McLellan retweeted
Same
@elonmusk @TeslaCharging , Can we get an off-road Supercharger?
Our current setup is a bit dated:
Diesel → Generator → 480VAC → 160kW DC → Cybertruck.
QP/share whatever you like for #WhateverWednesday
#5:
"You need to open up cloudflare and..."
"I have the cloudflare connector"
"You're right!"
teapawt.net/ChadMoran/status/21152…
What if the opening sequence was OpenAI's own world model?
