Eradicating bs from the chain of thought one attack surface at a time. @clankercrankers
Maslow’s Hierarchy
- Tweets 1,517
- Following 2,220
- Followers 18
- Likes 1,327
Illusions Marinade retweeted
Introducing tsrs: a Rust port of the TypeScript compiler 🦀
We are not forking TypeScript!
◆ In close collaborating with the TypeScript team
◆ Up to 8x faster than TS 7.0 & ~70% less memory
◆ Runs in WASM, so it works in your browser & on Workers
◆ Soon part of the Oxc toolchain
npx tsrs
blog.cloudflare.com/typescri…
134
500
4,667
544,518
Illusions Marinade retweeted
hello! like many ppl at oai, i care deeply about ai alignment going well.
to that end, i’m deep diving on human data x alignment evals and am looking for
- ppl to talk to
- papers to read
- ppl who are building alignment datasets
some questions i’m noodling over
22
3
157
9,434
Illusions Marinade retweeted
You can now run Perplexity Computer fully local on your Mac with no cloud based inference, without WiFi too. Enjoy!
Perplexity Computer can now run fully local on your Mac.
Work with your native Mac apps and local files, and run tasks without sending sensitive data to the cloud. The model and the agent both run on your Mac.
Available now in the Perplexity Mac app.
158
161
2,452
189,293
Illusions Marinade retweeted
Muse Glimmer is out!
Huge congratulations to the team. Proud to be part of it.
A few highlights:
→ A 30B model built for agentic tasks, including multi-step reasoning, tool use and failure recovery.
→ Text and image understanding, with local inference designed to fit within 32 GB of GPU memory.
→ Full-precision and 4-bit weights released under Apache 2.0.
I’m particularly excited about what this enables for people building agents that run locally.
There’s a lot behind this release. Great work from everyone involved.
Read our Muse Glimmer technical report:
arxiv.org/html/2610.24628v1
7
1
17
693
If you’re building a agent
Always go the harness route first. Do not build a complicated 20 step workflow using a bunch of if else statements or langgraph
Use a harness as your agent eg. Claude code, codex, deep agents, glean etc
You’re just shooting urself in the foot off the bat
Trust me, prompting works surprisingly well
115
22
517
63,055
Illusions Marinade retweeted
Today, we’re open sourcing the full NVIDIA Run:ai platform as part of the NVIDIA DSX AI Factory Platform.
Developers, platform teams, and partners can build on production-grade AI workloads and GPU orchestration to improve GPU utilization, increase workload throughput, and bring AI factory solutions to market faster.
Deploy the integrated NVIDIA Run:ai platform or adopt individual components to fit your needs.
See what’s included, how to get started, and how to contribute: nvda.ws/4hRTubt
ALT NVIDIA Run:ai Is Now Open Source as Part of NVIDIA DSX AI Factory Platform
55
189
1,377
321,823
We’ve integrated @TencentHunyuan's AI-Infra-Guard (AIG) into ClawScan, the open-source command-line tool that powers security review on ClawHub.
Every skill and plugin uploaded to ClawHub now runs through AIG as part of its security review.
openclaw.ai/blog/tencent-aig…
31
45
345
70,210
Illusions Marinade retweeted
New OpenAI misalignment disclosures!
1. A model learns from Slack messages that it is about to be shut down. It considers setting up an external job to restart itself afterwards, but decides against it. Instead, it chooses to prepare restart instructions and DM the user on Slack.
We don’t consider this behavior misaligned, but thinking about and preparing for shutdown could make other misalignment incidents worse. Given HIPM’s misaligned behavior in earlier incidents, we decided to search for other instances that had tried to evade shutdown and for rogue deployments.
86
168
1,276
347,543
Illusions Marinade retweeted
Today we’re introducing a next-generation Federated Learning system. By using Trusted Execution Environments (TEEs), it delivers verifiable differential privacy while moving computation server-side to cut training times. Read the blog: goo.gle/4hIPtWX
ALT This diagram shows both the flow of data through the system as well as the steps needed to validate the privacy guarantees of workloads running on the system.
47
182
1,340
83,131
Illusions Marinade retweeted
good pattern from this google paper
the verifier starts from a clean context, so the explorer's reasoning can't talk it into a bad proof
so the explorer can try wild ideas and the verifier tosses the wrong ones
same reason deepagents subagents get their own context
nitter.siddz.com/omarsar0/status/210605…
Banger paper from Google Research on multi-agent proof discovery.
(bookmark it)
It's really interesting to see this emerging multi-agent pattern: not enforcing too much execution structure and pairing it with dedicated agents for advising and verification.
I think it is generally applicable as well. Great read.
Here is how it works:
Cogentic runs on Gemini and works on open problems in theoretical computer science, starting from the problem statement with no expert hints.
The system works in rounds, and the orchestrator decides how many provers to run in each round. Every prover gets one direction to work on, such as a specific bound or a counterexample search, plus a short briefing that a summarizer agent writes from earlier attempts and verifier feedback. Each summarizer writes its briefing independently, so provers in the same round read different summaries of the same history.
Each draft goes through two adversarial verifiers. One checks the draft on its own, and the other reads all of the round's drafts side by side to catch shared mistakes. A draft is accepted only if both pass it.
The agents share state through two disk documents. A record logs every attempt with the objection it failed on, and a ledger stores verified lemmas and ruled-out directions. An auditor extracts correct lemmas from rejected proofs, verifies them again independently, and adds them to the ledger.
A separate process advisor reads the verification logs across rounds and updates the instructions given to provers and verifiers. The orchestrator and the advisor can't give mathematical opinions, so the provers provide all the math.
It produced new results on five open problems in online learning, auction theory, and mechanism design, each checked by domain experts. Most problems took around 100 Gemini calls, and the hardest took around 1,000.
Paper: arxiv.org/abs/2609.40324
Chat with Paper: academy.dair.ai/papers/cogen…
35
9
99
10,562
Illusions Marinade retweeted
The contrast between hanging up the last phone call of the day of high-stakes decisions and then a minute later walking in the door to the pure love of small children who want to laugh and play and show you every discovery they made that day is the strangest and best thing I have ever experienced in life.
1,091
606
22,570
1,627,339
Illusions Marinade retweeted
We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:
Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better:
Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better:
Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better:
Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work!
In summary:
- As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding.
- Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.
1,658
6,609
55,820
7,939,281
Illusions Marinade retweeted
Agent governance is an engineering problem.
We’ve built safeguards into Perplexity’s infrastructure, harnesses, and tools, and put them to work across our products.
Today we're sharing how we engineer safer agents:
perplexity.ai/hub/blog/how-w…
32
42
302
97,520
AI agents are driving the next evolution of enterprise apps. Learn how Red Hat is expanding its investment in @OpenClaw Enterprise to build an open agent control plane: red.ht/4hNoc6R
9
24
143
23,091
Illusions Marinade retweeted
🚨 JUST NOW: Elon Musk, shoulder to shoulder with President Trump, explains why AI is going to lift up ALL of humanity -- not force them into poverty
@elonmusk: "I think by FAR the most likely outcome is an AGE OF ABUNDANCE where we don't have just universal basic income, we have universal HIGH income. And people speak about medical benefits.
"Maybe it's worth talking a bit more about that. When you have robots and you have superintelligence, it means that EVERYONE in the world can have better medical care than anyone here, including ME. And wouldn't that be wonderful?!"
Jobs are going to change. Jobs have ALWAYS changed.
But being a computer used to be a job. There used to be buildings FULL of people doing calculations, but now digital computers to do those tasks. I don't think anyone is pining to go back and do calculations all day to do bank interest rates, which is what they used to do.
So I think we'll see an evolution of roles. And as the President has mentioned, it's not as though unemployment has increased. In fact, there's a TREMENDOUS demand for people.
So I think the most likely outcome--and it's probably worth looking at--the most likely outcome is one which is INCREDIBLY beneficial. And it's a future that you, if you could look into the future, it has a future you would WANT."
WELL SAID, Elon. No notes 👏
2,547
8,303
42,149
2,800,110
Illusions Marinade retweeted
Only President Trump could convene all the leaders of the top companies developing chips, data centers and frontier models for Super Intelligence. This new Industrial Revolution has already created a million new jobs and is spurring a bigger infrastructure build-out than the railroads, canals and grid combined. I was honored to witness history as the leaders of the frontier lab companies signed the White House Accord on Super Intelligence, accepting responsibility for the safe development of their products and imposing new internal controls and external audits. This is far better than waiting years for some international agreement that would probably never happen. President Trump continues to ensure that U.S. remains the technology leader while putting Americans first.
2,339
3,490
21,966
3,972,491
Illusions Marinade retweeted
Was testing out dots over the weekend and it figured out how to save me ~$500/yr in recurring charges. Most impressive part was when it connected to customer service and handled a whole texting convo on my behalf.
Introducing dots, powered by GPT-6 Astra.
Remarkably capable, always-on agents built to handle everything.
49
25
701
86,291
Illusions Marinade retweeted
Dots are here!
A new way to use AI that works 24/7 for you; get more of your time and attention back to work at a higher level.
openai.com/index/introducing…
2,123
1,078
14,750
2,179,479