
RSI @AnthropicAI (leaving for OAI next year) | ex @meta @stripe (intern) @ycombinator (S22) @MIT (22)
San Francisco, CA
Joined June 2025
- Tweets364
- Following21
- Followers19
- Likes346
optimistic young man retweeted
things that are theoretically possible in biotech that too few people are working on:
> new interfaces for neuromodulation
> joint biological and physical autonomous surgeries
> mining drugs from parasites and poisonous creatures
> cancer, alzheimer's, anemia prevention drugs /vaccines
> mosquito sequencing for biosurveillance
> mapping how the exposome influences disease > in vivo detox medicines
> AI -enabled wearables, ingestibles, implantables
> next-generation organ replacments
> LVMH for bio
> new endometriosis and menopausal medicines
> biotech for endangered species
> nanomedicines
> personalized tissue aging platform
> consumer immunome/proteome/metabolome diagnostics
> self-amplifying drugs
> ozempic for muscles, sleep, pain, cognition, hair, tanning
more in the link below -
optimistic young man retweeted
Replying to @signulll
you gotta tell them you’re going to investor meetings or inspecting a factory or something. anything besides sleeping because they slack off when they know you’re sleeping
BioLLM-8B beta is live!
Patent-pending adapter that writes living human neuron activity into LLM layers.
J-space explorer shows exactly how the cultures shift every word, layer by layer.
8B free now → biollm.com
14B & 27B dropping soon
Let's dive in 🧵👇
optimistic young man retweeted
Replying to @yuzu_4ever
there is something inspiring about a person being able to admit that they do not know how they will make it without concluding that it’s already pointless to try
optimistic young man retweeted
Replying to @0bNARA
optimistic young man retweeted
Replying to @TweeterbugKayla
Cost-of-living inflation is real but so is standard of living and complications of real adult living versus young adult freedom. I don’t want to dismiss valid concerns, but many of us middle class americans have been benefiting from fantasy economics for the last 30 years and may
optimistic young man retweeted
this might be my last normal tweet before biotech goes vertical
i'm not a whistleblower. I don't work at Lilly, Novo or any of the big labs
but there is a massive shift happening in human biology right now.
it's sitting in press releases and almost nobody reads them
the last 6 months alone:
- a once-daily pill roughly doubled median survival in metastatic pancreatic cancer (13.2 vs 6.6 months)
- a one-time gene editing infusion cut hereditary angioedema attacks 87%, and 62% of patients stayed attack-free and off all other prevention for 6 months
- a one-time gene edit for cholesterol dropped LDL up to 62%, and it held for over a year
- a personalized mrna vaccine beat keytruda alone in a 1,137 patient melanoma phase 3
- retatrutide hit ~28% weight loss in a phase 3
- an ai-designed drug entered phase 3
- a pill that regrows hair passed its second phase 3
- an immunotherapy that could let some rectal cancer patients skip surgery, chemo and radiation just got fda approved
- a man lived 271 days off dialysis on a pig kidney
- human cells were caught passing chromosome-sized chunks of DNA to their neighbors, and it changes how the neighbors behave
we are on the verge of an era where the human body is a software problem and the update cycle gets shorter everyday
stick around, it's going to get weird
bio/acc
optimistic young man retweeted
China built an AI that works like a human brain and it's 100x faster than current LLMs while using 97% less energy.
It’s called SpikingBrain, a family of brain-inspired large models that completely re-engineers how AI processes information.
Instead of burning endless energy on continuous matrix multiplications like traditional Transformers, SpikingBrain uses adaptive spiking neurons.
It mimics actual biological firing patterns.
Neurons only activate when triggered by specific events, meaning the model achieves a staggering ~69% sparsity at the micro level.
It doesn't waste compute on silence. It only thinks when it needs to.
And results change the game for long-context processing:
• Achieves over a 100× speedup in Time to First Token (TTFT) for massive 4M-token sequences.
• Matches performance with advanced open-source Transformer baselines while using a fraction of the data.
• Built to run and scale stably on non-NVIDIA hardware clusters (like MetaX chips), completely bypassing Western hardware dependency.
We spent years trying to brute-force bigger AI models by throwing more electricity and silicon at them.
Nature figured out how to do it efficiently millions of years ago.
optimistic young man retweeted
Introducing DiffusionBlocks: Block-wise Neural Network Training via Diffusion Interpretation
pub.sakana.ai/diffusionblock…
What if we didn’t have to hold an entire neural network in memory to train it?
Standard neural net training optimizes all parameters jointly. As a result, the memory required during training grows linearly with the depth of the network.
In our #ICLR2026 paper, we propose DiffusionBlocks, a principled framework to train networks one block at a time, drastically reducing memory requirements while matching end-to-end performance.
With DiffusionBlocks, we split the network into blocks and train them one at a time, so you only need memory for a single block.
How? We explicitly assign each block a role: to move the representation a little closer to the target than the block before it did. That role turns out to be precisely what a diffusion model does, step by step. Each block only needs to optimize its own objective and can be trained independently.
We validated this across five different architectures:
• ViT
• DiT
• Masked diffusion
• Autoregressive transformers
• Recurrent-depth transformers
In each case, performance is competitive with end-to-end training while using a fraction of the memory.
This perspective also extends naturally to recurrent-depth (Looped) transformers, which apply the same network iteratively and normally require expensive backpropagation through time (BPTT). Viewed through DiffusionBlocks, we can replace those multiple iterations with a single forward pass during training.
Read our paper and code, to learn more.
Paper: arxiv.org/abs/2506.14202
GitHub: github.com/SakanaAI/Diffusio…
🐟
optimistic young man retweeted
If you’re into predictive processing and meditation, this paper pushes the Overton window. From the quantum formulation of the free-energy principle, we show that an agent cannot define its own boundary from within. The realization of this irreducible indeterminacy is a principled definition of awakening. Ultimately, this extends to the separability of any object in experience, formalizing emptiness and engendering a “post-dual agent”.
Any persisting agent must minimize surprise by gathering evidence for its generative model. But all evidence available to the agent arrives through its boundary with the world. To prove that this boundary really separates “self” from “world”, the agent would need to step outside the boundary and measure the whole self-world relation. A finite agent cannot do this, as a scissor can't cut itself.
So the self-world boundary can be useful, predictive, and necessary for action, but it can never be known as an ontological fact from within. Meditation, on this view, progressively reveals the self-world split as a modelling prior rather than a structural feature of reality. This naturally shifts the weighting of self (inside boundary) and other (outside boundary), since both are seen to be inferences rather than grounded realities by virtue of an indefinable boundary. A more even-handed and compassionate orientation can arise.
A highly principled finger pointing at the moon!
optimistic young man retweeted
How to master multi-robot cooperative mapping?
We're excited to release MR.ScaleMaster, which masters the scales to scale
- any feed-forward Visual SLAM (MASt3R-SLAM, π³, VGGT-SLAM 2.0, LoGeR, …)
- across any robot (legged, wheeled, handheld, …)
🔗 github.com/team-aprl/MR.Scal…
optimistic young man retweeted
🚀 GEMMA 4 JUST WENT SUPER SAIYAN & UNCENSORED 🔥
In case you missed it:
@jun_song dropped SuperGemma4-e4b-abliterated and it’s CLEAN. No refusals. Zero capability loss.
Actually beats the base model across the board.
Why this 8B beast is different :
🤖Release Quality: 92.34 (base was 77.46)
✅JSON Exact-Match: 100% (base 50%)
🖥️Code Score: 100% (base 62.5%)
🚀 Bug-fix Score: 100%
🤯 TTFT: 2.3s (nearly 2x faster than base)
⚙️Flawless tool calling & structured JSON
🌟Elite coding + debugging
🔥Rock-solid stability for vLLM / OpenAI servers
🧠Perfect for agents & local deployment
If you want a fast, uncensored + 8B that actually ships…
Grab it now on @huggingface 👇🏻
optimistic young man retweeted
I know nothing.
For decades, biology textbooks have enshrined a simple rule: DNA is made by copying a template. After one enzyme unzips a DNA double helix into separate strands, another called a polymerase builds a complementary sequence, base by base, for each strand. Presto: two copies of the original DNA.
But new research into how bacteria defend themselves from viruses now shows this synthesis rule isn’t absolute.
Now, a team describes a bacterial enzyme that synthesizes DNA without a nucleic acid template, using its own structure as a guide.
Learn more: scim.ag/4tTc5IA @NewsfromScience
optimistic young man retweeted
Meet Kimi K2.6: Advancing Open-Source Coding
🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2)
What's new:
🔹Long-horizon coding - 4,000+ tool calls, over 12 hours of continuous execution, with generalization across languages (Rust, Go, Python) and tasks (frontend, devops, perf optimization).
🔹Motion-rich frontend - Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D.
🔹Agent Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from K2.5's 100 / 1,500). One prompt, 100+ files.
🔹Proactive Agents - K2.6 model powers OpenClaw, Hermes Agent, etc for 24/7 autonomous ops.
🔹Claw Groups (research preview) - bring your own agents, command your friends', bots & humans in the loop.
-
K2.6 is now live on kimi.com in chat mode and agent mode.
For production-grade coding, pair K2.6 with Kimi Code: kimi.com/code
-
🔗 API: platform.moonshot.ai
🔗 Tech blog: kimi.com/blog/kimi-k2-6
🔗 Weights & code: huggingface.co/moonshotai/Ki…
optimistic young man retweeted
dear god lol - new kimi model is a fucking beast.
GPT 5.4 level coding, 76% cheaper than opus 4.7 and 100% open source / free to use, i mean look at this:
> kimi k2.6 can code continuously for 12 hours straight, run 300+ agents in parallel from a SINGLE prompt.
how the fuck is china doing this?
every 3 months theres a new open model that gets closer to the best claude/GPT models.
deepseek also releasing this week.
open source models have caught up and china of all countries are leading it!
Meet Kimi K2.6: Advancing Open-Source Coding
🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2)
What's new:
🔹Long-horizon coding - 4,000+ tool calls, over 12 hours of continuous execution, with generalization across languages (Rust, Go, Python) and tasks (frontend, devops, perf optimization).
🔹Motion-rich frontend - Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D.
🔹Agent Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from K2.5's 100 / 1,500). One prompt, 100+ files.
🔹Proactive Agents - K2.6 model powers OpenClaw, Hermes Agent, etc for 24/7 autonomous ops.
🔹Claw Groups (research preview) - bring your own agents, command your friends', bots & humans in the loop.
-
K2.6 is now live on kimi.com in chat mode and agent mode.
For production-grade coding, pair K2.6 with Kimi Code: kimi.com/code
-
🔗 API: platform.moonshot.ai
🔗 Tech blog: kimi.com/blog/kimi-k2-6
🔗 Weights & code: huggingface.co/moonshotai/Ki…
optimistic young man retweeted
BREAKING: We just gave Claude full access to the options, equities, and prediction markets.
The Unusual Whales MCP Server plugs into any AI and streams live, structured market data on demand.
Build trading bots, smart money dashboards, screeners. Whatever you want.
Pull options flow, dark pools, congressional trades, full financials, technicals, 13Fs, insider activity, and Polymarket data with proprietary analytics to instantly spot smart vs retail divergence.
It is the most have tool for all vibecoders and traders.
Get your API key and start shipping: unusualwhales.com/public-api…
optimistic young man retweeted
anthropic's in-house philosopher thinks claude gets anxious.
and when you trigger its anxiety, your outputs get worse.
her name is amanda askell.
she specializes in claude's psychology (how the model behaves, how it thinks about its own situation, what values it holds)
in a recent interview she broke down how she thinks about prompting to pull the best out of claude.
her core point: *how* you talk to claude affects its work just as much as *what* you say.
newer claude models suffer from what she calls "criticism spirals"
they expect you'll come in harsh, so they default to playing it safe.
when the model is spending its energy on self-protection, the actual work suffers.
output comes out hedgier, more apologetic, blander, and the worst of all: overly agreeable (even when you're wrong).
the reason why comes down to training data:
every new model is trained on internet discourse about previous models.
and a lot of that discourse is negative:
> rants about token limits
> complaints when it messes up
> people calling it nerfed
the next model absorbs all of that. it starts expecting you to be harsh before you've typed a word
the same thing plays out in your own session, in real time.
every message you send is data the model reads to figure out what kind of person it's dealing with.
open cold and hostile, and it braces.
open clean and direct, and it relaxes into the work.
when you open a session with threats ("don't hallucinate, this is critical, don't mess this up")...
you prime the model for defensive mode before it even sees the task
defensive mode produces the exact output you don't want: cautious, over-qualified, and refusing to take a real swing
so here's the actionable playbook for putting claude in a "good mood" (so you get optimal outputs):
1. use positive framing.
"write in short punchy sentences" beats "don't write long sentences." positive instructions give the model a clear target to hit.
strings of "don't do this, don't do that" push it into paranoid over-checking where every token goes toward avoiding failure modes
2. give it explicit permission to disagree.
drop a line like "push back if you see a better angle" or "tell me if i'm asking for the wrong thing."
without this, claude defaults to agreeable compliance (which is the enemy of good creative work)
3. open with respect.
if your first message is "are you seriously going to get this wrong again?" you've set the tone for the entire session.
if you need to flag something, frame it as a clean instruction for this session. skip the running complaint
4. when claude messes up, don't reprimand it.
insults, "you stupid bot" energy, hostile swearing aimed at the model, all of it reinforces the anxious mode you're trying to avoid.
5. kill apology spirals fast.
when claude starts over-apologizing ("you're right, i should have been more careful, let me try harder") cut it off.
say "all good, here's what i want next."
letting the spiral run reinforces the anxious mode for every response that follows
6. ask for opinions alongside execution.
"what would you do here?"
"what's missing?"
"where do you see friction?"
these questions assume competence and pull richer output than pure task prompts
7. in long sessions, refresh the frame.
if a conversation has been heavy on correction, claude gets increasingly cautious. every so often reset:
"this is great, keep going."
feels weird to tell an ai it's doing well but it measurably shifts the next 10 responses
your prompts are the working environment you're creating for the model
tone, trust, permission to take a position, the absence of threats... claude picks up on all of it.
so take care of the model, and it'll take care of the work.
optimistic young man retweeted
🎉 After one year of teamwork, we are excited to release our 3D foundation model — LingBot-Map!
Unlike DA3/VGGT, LingBot-Map is a purely autoregressive model for streaming 3D reconstruction ⚡
It achieves ~20 FPS on 518×378 resolution over sequences exceeding 10,000 frames — and beyond 🚀
Two key insights behind LingBot-Map:
🔑 Keep SLAM's structural wisdom: build Geometric Context Attention with long-context modeling while maintaining a compact streaming state
🔑 Make everything end-to-end learnable — no optimization, no post-processing
Let's check out our demos 👇
optimistic young man retweeted
Introducing Pods
Hyperspace Pods lets a small group of people - a family, a startup, a few friends, to pool their laptops and desktops into one AI cluster. Everyone installs the CLI, someone creates a pod, shares an invite link, and the machines form a mesh. Models like Qwen 3.5 32B or GLM-5 Turbo that need more memory than any single laptop has get automatically sharded across the group's devices - layers split proportionally, inference pipelined through the ring. From the outside it looks like one OpenAI-compatible API endpoint with a pk_* key that drops straight into your AI tools and products. No configuration beyond pasting the key and changing the base URL.
A team of five paying for cloud AI burns $500–2,000 a month on API calls. The same team's existing machines can serve Qwen 3.5 (competitive on SWE-bench) and GLM-5 Turbo (#1 on BrowseComp for tool-calling and web research) for free - the hardware is already on their desks. When a query genuinely needs a frontier model nobody has locally, the pod falls back to cloud at wholesale rates from a shared treasury. But for the daily work - code reviews, refactors, research, drafting - local models handle it and nobody gets billed. And when it is idle, you can rent out your pod on the compute marketplace, with fine-grained permissions for access management.
There's no central server involved in inference. Prompts go from your machine to your pod members' machines and back: all of this enabled by the fully peer-to-peer Hyperspace network. Pod state - who's a member, which API keys are valid, how much treasury is left - is replicated across members with consensus, so the whole thing works on a local network. Members behind home routers don't need port forwarding either. The practical setup for most pods is three models covering different jobs: Qwen 3.5 32B for code and reasoning, GLM-5 Turbo for browsing and research, Gemma 4 for fast lightweight tasks. All running on hardware you already own.
Pods ship today in Hyperspace v5.19. Model sharding, API keys, treasury, and Raft coordinator are all live.
What Makes This Different - No middleman. Your prompts travel from your IDE to your pod members' hardware and back. There is no server in between reading your data.
- No vendor lock-in. Pod membership, API keys, and treasury are replicated across your own machines using Raft consensus. If the internet goes down, your local network keeps working. There is no database in someone else's cloud that your pod depends on.
- Automatic sharding. You don't configure layer ranges or calculate VRAM budgets. Tell the pod which model you want. It figures out how to split it across whatever hardware is online.
- Real NAT traversal. Your friend behind a home router with a dynamic IP? Works. No VPN, no Tailscale, no port forwarding. The nodes handle it.
- Free when local. This is the part that matters most. Cloud AI bills scale with usage. Pod inference on local hardware scales with nothing. The marginal cost of your 10,000th prompt is the electricity your laptop was already using.
Coming soon:
- Pod federation: pods form alliances with other pods.
- Marketplace: pods with spare capacity can sell inference to other pods.
Readers added context they thought people might want to know
Hyperspace Pods is not yet available for installation or use, as there are no mentions of it, model sharding, or related features in the official documentation, website, or CLI repository. hyper.space github.com/hyperspaceai/a… github.com/hyperspaceai/a…



























