mycall

11 ms·

151.
by mycall 8mo ago
Trees of trust, self-organizing as closer nodes to your agents are trusted more.
152.
by mycall 8mo ago
It corrupts future AI models too, so we might be holding onto today's models for a long time, as a least biased version as a checksum.
153.
by mycall 8mo ago
You never went to a LAN party then
154.
by mycall 8mo ago
That isn't good enough and could be DDoS'd as well.
155.
by mycall 8mo ago
Make it private invite only behind Cloudflare
156.
Neurophos raises $110M to build tiny optical processors for inferencing(techcrunch.com)
1 points by mycall 8mo ago|0 comments
157.
by mycall 8mo ago
Also, getting insurance to pay for cybercrimes is hard and sometimes doesn't justify their costs.
158.
by mycall 8mo ago
Perhaps those companies should have performed verified backups of third-party vendor's published security policies into a secure enclave with paired keys with the auditor, to keep a trail of custody.
159.
by mycall 8mo ago
Sucker?
160.
by mycall 8mo ago
I agree. Unless they fix things, this will crash and burn, but the idea still has a future.
161.
by mycall 8mo ago
That wrongness is the frontier labs trying to remove their benchmaxxing bias, so the models now have a concept of 'I don't know' and will rethink directions and goals better. There was lots of research last year on this topi
162.
by mycall 8mo ago
It doesn't take much effort to find news about AI (LLMs) successfully being deployed with ROI in healthcare, legal, customer operations, retail, banking accounting/tax and more. I don't think the article needs to worry about
163.
by mycall 8mo ago
Living in the US, my trust in US firms and state is also completely gone. My only hope is local LLMs.
164.
by mycall 8mo ago
It's in chicken-egg mode, where could be useful if more people and bots used it, but not there yet.
165.
by mycall 8mo ago
At least with GPT-5.3-Codex-Spark, I gather most of the AI inference isn't rendering cat videos but mostly useful work.. so I don't feel tooo bad about 16 households worth of energy.
166.
by mycall 8mo ago
If I remember, both Claude Code and OpenAI Codex "harnesses" improved themselves now. OpenAI used early versions of GPT-5.3-Codex to: debug its own training process, manage its deployment and scaling and diagnose test results and
167.
by mycall 8mo ago
- Tinkering: gpd win 5 (112GB VRAM) + Agent-Zero + Gwen3-Coder-Next-Q8_0 + Unreal Engine 5 + gnuradio + b206mini = decode/encode 4D visualizations with context understanding by data frame captures/generation through gnuradio self-
168.
by mycall 8mo ago
That was 10 days ago. I wonder if the discussions the moltys have begin to converge into a unified voice or if they diverge into chaos without purpose.
169.
by mycall 8mo ago
Until use of TCP/IP or DNS requires uploading a video scan, the internet will keep on.
170.
by mycall 8mo ago
> Your server needs to authenticate itself with Apple, and this requires an Apple account How does Firebase Cloud Messaging work with Apple without an Apple account, or is that implied in the client generated push token residing in Fireb
171.
by mycall 8mo ago
In the age of OpenClaw and interest agentic processes, the LLM doesn't need to dream up everything by itself. There are tons of tools it can use. Using those tools, the LLM can generate a random world and write about that, and score
172.
by mycall 8mo ago
What does ANTHROPIC bring to this project that a local LLM cannot, e.g. Gwen3 Coder Next?
173.
by mycall 8mo ago
> Atari’s primary coin-op manufacturing and headquarters facility was located at 1265 Borregas Ave, Sunnyvale, CA 94089 during the key 1976–1984 period. Another critical site in the immediate area was 1196 Borregas Avenue, used during th
174.
by mycall 8mo ago
Same reason people build their own homes by hand, for the challenge because YOLO.
175.
by mycall 8mo ago
We are becoming spec writers, wearing the PM/lead hats. 1) Do a gap and needs assessment. 2) Build business requirements. 3) Define scope of work to advance fulfillment. 4) Create functional and non-functional specs. 5) Divide-conquer-
176.
by mycall 8mo ago
Isn't that what Tolkien did in his head? Write something, learn what he liked/didn't like then revise the words? Rinse/repeat. Same process here.
177.
by mycall 8mo ago
That is still important for safety reasons in case someone uses a LiDAR jamming system to try to force you into an accident.
178.
by mycall 8mo ago
There is tons of improvements in the near future. Even Claude Code developer said he aimed at delivering a product that was built for future models he betted would improve enough to fulfill his assumptions. Parallel vLLM MoE local LLMs on
179.
by mycall 8mo ago
Why not do a load balanced approach two multiple models in the same chat session? As long as they both know each exists and the pattern, they could optimize their abilities on their own, playing off each other's strengths.
180.
by mycall 8mo ago
Now I like that idea as an environment variable that takes precedence over the command parameters.

More ›