These guys are artists. Moment Motor takes classic cars and turns them electric. It's two to four months of work and can cost $175,000. Tons of people in tech land send their cars to Austin to get modernized.
Full episode here.
Weekend experiments: got local TTS running on mobile with QVAC, and the results are looking really promising. It's great to see more AI workloads moving fully on-device.
New article dropping soon with a deeper dive. Mobile inferencing is coming along nicely. ๐
@qvac
You do realize thereโs no AI bubble and weโre not going back to writing code by hand, ever?
Hint: thereโs a reason why youโre not writing assembly by hand
When if you're reviewing a PR that was mostly done by AI, you ask the PR author a question, they can't explain what or why they this or that! ๐ญ
They literally prompt back into AI chatbot again and come back with another AI slop response. This here is even more sad
Something very sad about devs who are now fully relying on AI code generation is the fact that they can't really think for themselves anymore.
This is literally brain atrophy happening in real time. Very sad!
Some lore on mobile inferencing:
Iโve been playing around with QVAC lately, and Iโm honestly surprised by how fast local inference has become on modern phones.
A few years ago, the idea of running an LLM directly on a smartphone sounded ridiculous.
@qvac
We are investigating unauthorized access to GitHubโs internal repositories. While we currently have no evidence of impact to customer information stored outside of GitHubโs internal repositories (such as our customersโ enterprises, organizations, and repositories), we are closely monitoring our infrastructure for follow-on activity.
customers are increasingly asking us for certainty on capacity. as models get better, we expect that the world will be capacity-constrained for some time.
we are offering discounted tokens for 1-3 year commits.
(it also helps us plan, so hopefully a big win-win.)
It is a shame that the simple act of transferring a large block of data as fast as possible over the internet is not handled effectively by the primitive operating system calls. You either multiplex over parallel persistent TCP connections to combat head-of-line blocking and slow starts, or reinvent reliable delivery and flow control over UDP.
QUIC has a lot going for it, but it is a large library (six figure LoC!) and conflates security and performance in a way I donโt love. There is also fundamental information about competition with other processes and link layer congestion that should be useful, but is unavailable to user libraries.
You should be able to just write(really_big_buffer) and it is all taken care of for you.
Microsoft is investigating mistralai PyPI package v2.4.6 compromise. Attackers injected code in mistralai/client/__init__.py that executes on import, downloads hxxps://83[.]142[.]209[.]194/transformers.pyz to /tmp/transformers.pyz, and launches a second-stage payload on Linux. The file name transformers.pyz appears deliberately chosen to mimic the widely used Hugging Face Transformers library and blend into ML/dev environments.
The main payload is a credential stealer, but it also includes country-aware logic; it avoids Russian-language environments and contains a geo fenced destructive branch that has 1-in-6 chance of executing rm -rf / when the system appears to be in Israel or Iran.
To mitigate this threat: isolate affected Linux hosts, block 83[.]142[.]209[.]194, hunt for /tmp/transformers.pyz, pgmonitor[.]py, and pgsql-monitor.service, and rotate exposed credentials.
@weskambale This is not a JS/TS specific issue. Even python ecosystem is also affected by supply chain attacks.
It's just a by product of the current software development model.
This same attack today affected a Mistral python package as well