Co-founder of @open_rpc, prev: @etc_core, @consensys, @cellarius2084 and others... | Open Source acolyte
Joined August 2011
- Tweets887
- Following1.3K
- Followers402
- Likes2.5K
seeing all this jev @typesafeai stuff makes it feel as though it's kind of a static world model where there's some probability given to each prediction, I wonder if there's a way to make this smaller and realtime update weights to shift probabilities...
Pretty interesting, it's kind of an experiment in visualizing qwen's "perception of the internet" as determined by consumed content ๐ค
you can just hallucinate the entire internet with Qwen 3.8 27b running at 2,000 tokens/second?
part 2 of turning @cerebras + @Alibaba_Qwen 3.8 27B into an OS:
built an offline browser with zero network calls and mounted it directly the JIT ubuntu desktop.
no wifi. no scraping. zero packets sent to external CDNs.
you search a site, set a year, and qwen 27b at 1,950 tok/s synthesizes the entire DOM on the fly.
here is youtube in 2045 vs 1999:
โ search google for youtube inside the OS
โ scrub to 2045: instant futuristic feed
โ scrub to 1999: raw web 1.0 time capsule in seconds
at this speed, browsing isn't retrieving files from a server, it's querying an alternate reality.
a 2D browser window is just step zero.
imagine full operating systems, virtual worlds, and complex simulation engines existing purely as model weights. zero gigabytes stored on disk, just pure interactive reality streamed on demand.
What else becomes a possibility with the qwen 3.8 27b (dense) at 2000 tokens/sec?
Was talkin w/ @shanejonas, astra gens a ton of tests. My approx. for program complexity is 2^num_options, you prolly need 2^options test. It's too many. A sol. just have fuzzing tests and critical test. Don't review fuzz test, keep them, they're good to have, review crit test.
If we think about it critically, the tests that expand out to test the limits of say a number range or whitespaces or utf-8 vs codepoint are all tests we should have written, but would have never had time for pre AI. So let's not spend inf. hours trying to review those tests.
Zane Starr ๐ retweeted
Excited to announce that weโve received a grant from @GCCofCommons !
This grant will help us improve our tooling and deliver a better dev ex. for teams building with JSON-RPC.
Zane Starr ๐ retweeted
FAM!
Next month, We're hosting a FREE community family fishing day in my local community!๐ฃ
Over 250+ kids and their families signed up
And I need your support AND help.
We are about $6k short (due to funding cuts) and still need about 140 fishing poles/tackle boxes.
Links below...
Zane Starr ๐ retweeted
Software horror: litellm PyPI supply chain attack.
Simple `pip install litellm` was enough to exfiltrate SSH keys, AWS/GCP/Azure creds, Kubernetes configs, git credentials, env vars (all your API keys), shell history, crypto wallets, SSL private keys, CI/CD secrets, database passwords.
LiteLLM itself has 97 million downloads per month which is already terrible, but much worse, the contagion spreads to any project that depends on litellm. For example, if you did `pip install dspy` (which depended on litellm>=1.64.0), you'd also be pwnd. Same for any other large project that depended on litellm.
Afaict the poisoned version was up for only less than ~1 hour. The attack had a bug which led to its discovery - Callum McMahon was using an MCP plugin inside Cursor that pulled in litellm as a transitive dependency. When litellm 1.82.8 installed, their machine ran out of RAM and crashed. So if the attacker didn't vibe code this attack it could have been undetected for many days or weeks.
Supply chain attacks like this are basically the scariest thing imaginable in modern software. Every time you install any depedency you could be pulling in a poisoned package anywhere deep inside its entire depedency tree. This is especially risky with large projects that might have lots and lots of dependencies. The credentials that do get stolen in each attack can then be used to take over more accounts and compromise more packages.
Classical software engineering would have you believe that dependencies are good (we're building pyramids from bricks), but imo this has to be re-evaluated, and it's why I've been so growingly averse to them, preferring to use LLMs to "yoink" functionality when it's simple enough and possible.
This is an interesting take on the various crypto legislation being passed and how the future of the dollar is changing
The secret fight over Dollar 3.0 is happening RIGHT NOW in Congress.
Most people will sleep through itโฆ
and wake up in 2035 wondering why their money works completely differently.
Hereโs exactly whatโs on the table. โ
Dollar 1.0 = gold & silver
Dollar 2.0 = bank IOUs (your checking account)
Dollar 3.0 = tokenized, always-on, borderless USD
Weโre deciding its final shape in the next 6โ12 months.
This is bigger than the entire @federalreserve put together.
1/10 ๐งต
Zane Starr ๐ retweeted
1/ ๐ Grant Announcement: OpenRPC!
Co-funding a 6-month sprint to update @open_rpc, making it easier for developers, tooling, and infrastructure providers to build on Ethereum with consistent, open standards.
open-rpc.org/
I think this ๐ฏ , I think crypto stays insular and doesn't reach, because of the culture and this docuseries can help change that. I'd be excited to watch it!
If we are to onboard and welcome the next billion participants to @ethereum, we need to produce content that is ๐๐พ
free of jargon
amplifies diverse voices/experiences
meets ppl where theyโre at
moves away from hype x pseudo celeb driven content
is entertaining and
accessible
Zane Starr ๐ retweeted
@OctantApp opened its allocation window for epoch 9 last week
Please consider allocating to projects like mine that will interface with local communities, amplify builders, and tell the story of Ethereumโs communities across the globe
Learn more here: octant.app/project/9/0xca854โฆ
Creating benchmarks for multiturn data is a bit tricky, I find myself thinking about data in โช, working from the result to question or entrypoint, to then try and train the llm to go in the forward direction ๐
โก๏ธwhen we gen. code w/ prompts we're, really creating iterative prompts that work around the limits of transformer attention, by storing context out of band in the form of code edits. Then we prompt again with a fresh improved context via static analysis.
Waging "war" on our own people in LA and other cities with ICE, and then bombing Iran without cause, this administration can do better, the US and world deserve better than this.
I like the idea of this but for code, and maybe you could take incremental commits to build a world model for code. Then we wouldn't need to bootleg it each time as we prompt in claude/gpt etc... . It would be able to predict the future impact of code changes aka architect
Introducing V-JEPA 2, a new world model with state-of-the-art performance in visual understanding and prediction.
V-JEPA 2 can enable zero-shot planning in robotsโallowing them to plan and execute tasks in unfamiliar environments.
Download V-JEPA 2 and read our research paper โก๏ธ ai.meta.com/vjepa/
๐Super cool donation opportunity, a friend of mine is going to the Derby World Cup with an Indigenous Rising, a flat track roller derby team representing Native America, First Nations, and Polynesia gofundme.com/f/help-kristinaโฆ
๐ check it out pretty dope, could use it to preload an execution env. so an llm can generate code to interact in it. Offhand it could do interesting things in a custom user based env. A way to give llms async worker lvl access in a sandbox, it could read s3 direct even...
I made an Open Source MCP Server to run javascript inside a sandbox .
It uses SES (Secure Ecmascript) you haven't seen it, an Apache 2.0 licensed javascript sandbox from agoric.
It is also the underlying tech to securely sandbox metamask snaps.
github.com/shanejonas/javascโฆ