thomas_hnsn (@CorpReachNow) header
thomas_hnsn (@CorpReachNow) profile picture
Joined January 2022
Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)Photo by thomas_hnsn (@CorpReachNow)
thomas_hnsn retweeted
Been testing GPT 5.5 and it's implemented the HTML-in-Canvas spec Worked for both WebGL and WebGPU out of the box Then created this demo with the PixiJS + GSAP Skills Hopefully there will soon be no excuses to make boring websites
46
264
11
3,378
202,910
thomas_hnsn retweeted
the cult meets at midnight around the sacred RUST altar 137k lines. MIT licensed. zero isolation forbidden. 🐍 @openfangg
5
4
1
24
2,271
thomas_hnsn retweeted
I just noticed that someone created a vamp and copied my post to use as their pinned message in their community. Please be aware: this is not affiliated with us, not related to the original OpenFang discussion, and it is very likely a bundled launch. Do not trust it. Do not buy it. Always double check the exact name, CA, and source before interacting with any token.
3
2
1
17
1,493
thomas_hnsn retweeted
Quick update for the community. The thread I wrote about OpenFang including the breakdown and the CA in the comments just got reposted. I didn’t spam them. I simply shared a structured perspective, So that they don’t feel afraid or attacked. I’ve also sent a DM to Jaber and the OpenFang team to let them know that funding is currently available through the GitHub mechanism tied to $OpenFang.
5
5
16
1,814
Cute meme vibe, solid art, easy narrative to push. Most importantly — still low market cap, huge upside room. At this MC, it only takes a bit of FOMO volume to send the chart vertical.
16
2
5
2,261
thomas_hnsn retweeted
The AI agent space is maturing fast, and most people still don’t realize what’s actually changing under the hood. Let’s talk about @openfangg. ( The Agent Operating System ) OpenFang is not positioning itself as another chat-based agent framework. It introduces a fundamentally different approach: an Agent Operating System built from scratch in pure Rust, with 137,000+ lines of code, 1,700+ tests, MIT licensing, and zero Clippy warnings. Instead of wrapping tools around an LLM and relying on prompt chains, OpenFang runs agents inside WASM sandboxes under a kernel-like runtime. Each agent is scheduled, isolated, fuel metered, epoch controlled, and can be terminated instantly if it behaves unexpectedly. This treats agents as real system processes not experimental scripts. Security is a first class concern. @openfangg integrates 16 independent security layers, including dual metered WASM execution, Ed25519 signed manifests, Merkle hash chain audit trails, secret taint tracking with zeroization, prompt injection detection, SSRF protection, HMAC mutual authentication, rate limiting, subprocess isolation, and path traversal mitigation. Performance is equally notable. With a ~32MB binary and ~180ms cold start time, OpenFang maintains a lightweight footprint compared to heavier multi hundred megabyte stacks. OpenFang’s core positioning 7 autonomous Hands, 30 agent templates, 40 channel adapters, 38+ tools, 26+ LLM providers, 16 security systems, all inside a single binary. This is not a wrapper framework. It’s infrastructure. ( Comparison @openfangg & @openclaw ) Now, why does this matter especially in comparison to @openclaw? OpenClaw helped popularize agent tooling by making chat integrated workflows simple and accessible. It is flexible, model agnostic, and easy to deploy. However, its architecture remains largely prompt driven and reactive. Agents depend heavily on user interaction, and isolation is lighter compared to kernel-level sandboxing. The install footprint is significantly larger, cold starts are slower, and runtime control is less deterministic. OpenFang approaches the problem differently: • Kernel-level WASM isolation instead of loose tool execution • Deterministic scheduling and process control • Built in resource metering (fuel + epochs) • Stronger security model by default • Smaller binary footprint and faster startup • Designed for autonomous execution, not just chat response ( OpenFang vs The Landscape ) : This table visually shows the architectural differences. Rust vs TypeScript/Python 16 security layers vs 3 WASM sandbox vs minimal isolation Merkle chain audit trail vs basic logging 40 channel adapters vs 13 This isn’t about attacking competitors. It highlights a shift from convenience first frameworks to runtime first engineering. ( Hands + v0.1.1 ) The biggest shift is Hands, autonomous agents that operate continuously without constant prompting. Hands run on schedules, generate research, build knowledge graphs, automate workflows, monitor data streams, and deliver output directly to communication platforms. This transforms agents from reactive assistants into proactive AI workers. OpenFang v0.1.1 reinforces this production focused direction: • 30 agent templates bundled directly in the binary • Default provider consistency across agents • API keys persist across restarts • New providers added: Kimi, Qwen, MiniMax, Zhipu, Qianfan • OpenSSL dependency removed in favor of pure Rust TLS • Ubuntu 22.04+ Linux compatibility confirmed • Skill TOML parsing fixed • Docker image naming corrected These are infrastructure level refinements, not cosmetic updates. ( Benchmarks ) Cold start: ~180ms vs ~5960ms Idle memory: ~40MB vs ~394MB Install size: ~32MB vs ~500MB Security systems: 16 vs 3 These numbers reinforce the architectural intent: OS level determinism and lightweight execution. #AIAgents #OpenFang #OpenClaw #Rust #WASM #OpenSource
We open sourced an operating system for ai agents 137k lines of rust, MIT licensed we love @openclaw and it inspired a lot of what we built. but we wanted something that works at the kernel level so we built @openfangg agents run inside WASM sandboxes the same way processes run on linux. the kernel schedules them, isolates them, meters their resources, and kills them if they go rogue. it has 16 security layers baked into the core. WASM sandboxing, merkle hash-chain audit trails, taint tracking on secrets, signed agent manifests, prompt injection detection, SSRF protection, and more. every layer works independently. giving an LLM tools with zero isolation is insane and we're not doing it. we also created something called Hands. right now every ai agent is a chatbot that waits for you to type. Hands are different. you activate one and it runs on a schedule, 24/7, no prompting needed. your Lead Hand finds and scores prospects every morning and delivers them to your telegram before you wake up. your Researcher Hand writes cited reports while you sleep. your Collector Hand monitors targets and builds knowledge graphs continuously. they work for you. you don't babysit them github.com/RightNow-AI/openf… ⭐
2
6
31
4,340
thomas_hnsn retweeted
Im glad I checked both CAs, almost got rekt by the vamp haha. This looks good
3
1
1
19
2,815
thomas_hnsn retweeted
We have retardio, useless, pump, spx, fwog, troll and more whales positioning 😅
3
5
22
903
thomas_hnsn retweeted
308k views in 21hrs + 1.8k stars , 160 forks on github 🐍🐍🐍🐍 github.com/RightNow-AI/openf…
We open sourced an operating system for ai agents 137k lines of rust, MIT licensed we love @openclaw and it inspired a lot of what we built. but we wanted something that works at the kernel level so we built @openfangg agents run inside WASM sandboxes the same way processes run on linux. the kernel schedules them, isolates them, meters their resources, and kills them if they go rogue. it has 16 security layers baked into the core. WASM sandboxing, merkle hash-chain audit trails, taint tracking on secrets, signed agent manifests, prompt injection detection, SSRF protection, and more. every layer works independently. giving an LLM tools with zero isolation is insane and we're not doing it. we also created something called Hands. right now every ai agent is a chatbot that waits for you to type. Hands are different. you activate one and it runs on a schedule, 24/7, no prompting needed. your Lead Hand finds and scores prospects every morning and delivers them to your telegram before you wake up. your Researcher Hand writes cited reports while you sleep. your Collector Hand monitors targets and builds knowledge graphs continuously. they work for you. you don't babysit them github.com/RightNow-AI/openf… ⭐
6
6
1
34
5,601
thomas_hnsn retweeted
I wanna share what I’ve seen about @openfangg so far. I know this might be technical and not everyone will fully read or understand the breakdown and that’s okay. OpenClaw is currently viral. It’s simple, integrates easily with chat apps, and is model agnostic. That’s why it gained traction fast. But technically, it’s heavier (~500MB install), slower to start (~6s cold start), more prompt dependent, and has lighter isolation in terms of security. @openfangg takes a different approach. It’s built in Rust and positions itself as an “Agent Operating System.” It runs with WASM sandboxing, kernel-level isolation, resource metering, and multiple security layers. It’s much lighter (~32MB), faster (<200ms cold start), and introduces autonomous “Hands” agents that can run on schedules without constant prompting. That shifts the model from reactive chatbot to proactive AI worker. Yes, @openfangg is still very early. It’s not as battle tested as OpenClaw yet. But the attention is growing fast. Earlier today @Akashi203 post had only a few thousand views now it’s at 151k+. That kind of organic acceleration usually means real interest from builders and developers. Now it comes down to @Akashi203 response. My hope is that he acknowledges and accepts the funding model so the fees can genuinely support @openfangg long term development together with the community. If builder alignment this could turn into something much bigger. @Akashi203 if you read this, the community created this coin on @Pumpfun with one intention, to support @openfangg development. This is not about speculation or pressure. The structure was intentionally designed so that all fees go directly to the project’s GitHub. You can read this guide from @Pumpfun : nitter.gq/pumpfun/status/2022029… 7Z294H4R6FRmfQwX4XS8XfVUEk4zCb5cxMr1taudpump And sorry if this feels like too much posting.
We open sourced an operating system for ai agents 137k lines of rust, MIT licensed we love @openclaw and it inspired a lot of what we built. but we wanted something that works at the kernel level so we built @openfangg agents run inside WASM sandboxes the same way processes run on linux. the kernel schedules them, isolates them, meters their resources, and kills them if they go rogue. it has 16 security layers baked into the core. WASM sandboxing, merkle hash-chain audit trails, taint tracking on secrets, signed agent manifests, prompt injection detection, SSRF protection, and more. every layer works independently. giving an LLM tools with zero isolation is insane and we're not doing it. we also created something called Hands. right now every ai agent is a chatbot that waits for you to type. Hands are different. you activate one and it runs on a schedule, 24/7, no prompting needed. your Lead Hand finds and scores prospects every morning and delivers them to your telegram before you wake up. your Researcher Hand writes cited reports while you sleep. your Collector Hand monitors targets and builds knowledge graphs continuously. they work for you. you don't babysit them github.com/RightNow-AI/openf… ⭐
17
14
6
83
46,330
thomas_hnsn retweeted
Big things happening today with @pepecoins getting listed on @krakenfx ! 🐸🐙 To celebrate, I drew and minted a new NFT directly from Pepepaint on pepecoin.io that depicts this moment exactly as it occured: "The Creation of the Pepening" opensea.io/item/ethereum/0x2…
10
12
2
68
5,850
thomas_hnsn retweeted
/PepeCoin_World_Order 🐸 $pepecoin x #kraken 🧑‍🚒
5
12
73
1,160
thomas_hnsn retweeted
The OG frog is now live on the OG exchange @krakenfx. PepeCoin 🤝 Kraken. Frogs are in control. pro.kraken.com/app/trade/pep…
22
63
213
6,835
thomas_hnsn retweeted
gm frens 🔊 sound on ! the kek.space community is hard at work blocking off roads and demanding the conening 😅
27
59
3
231
14,438
thomas_hnsn retweeted
Yes
It’s no longer left versus right. What we are facing now is good versus evil.
12,203
34,850
1,391
324,547
27,945,965
thomas_hnsn retweeted
Anyone who celebrates murder is evil and should be utterly condemned
There’s a new political divide forming: If you celebrate when an innocent person whose political views differ from you is assassinated you’re gone. Don’t care what that person’s political persuasion is or how strongly you disagree with them. Cant think of one counterexample.
22,459
52,894
2,570
448,800
44,569,276
Good men must die, but death can't kill their names.
1,968
44,578
2,930
182,077
thomas_hnsn retweeted
SNEAKO speaks on people celebrating the passing of Charlie Kirk "So many pro-Palestine people are saying that I shouldn't pray for him because you disagreed with him. What's wrong with you? You think God wants you to celebrate his death? You think God doesn't want you to pray for his soul? This is a father. The country is going to go on a more chaotic route because of his death. This is a stain on the legacy of this country."
847
2,190
227
41,257
3,081,896