CJ 🇺🇸 retweeted
DeepSeek V4 Pro is live on Chutes.
It's the big one from DeepSeek's V4 family, with 1.6T parameters and a 1M-token context window. Open weights, MIT licence.
It's strong on agent work. DeepSeek's published score on Terminal-Bench 2.1 is 87.9, ahead of Opus 4.8 at 85.0.
$0.85 per million input tokens, $2.65 out.
Try it here: chutes.ai/app/chute/chutes-d…
CJ 🇺🇸 retweeted
Chutes is very excited to be working with METANOVA LABS as they develop their Hit-to-Lead optimization agent! We fully intend to assist and support to this project in any way we can as it develops.
NOVA Hit-to-Lead Optimization Agent
In collaboration with @chutes_ai and @onepot_ai
We're developing a Hit-to-Lead optimization agent that integrates an algorithm developed by a NOVA Blueprint winner, extending NOVA's capabilities beyond hit discovery and into the iterative optimization of small molecules.
The system combines a domain-specialized AI agent with proprietary combinatorial chemical databases supplied by our partner Onepot AI, spanning billions of possible compounds, that are delivered within 5-7 business days.
Researchers will be able to guide the agent in developing, comparing and refining optimization strategies, exploring chemical modifications and prioritizing candidates based on predicted affinity, selectivity, physicochemical properties and specific binding interactions.
Confidentiality is a central design requirement. Pharmaceutical R&D involves valuable intellectual property, including compound structures, targets, emerging chemical series and optimization strategies that cannot be exposed publicly.
The Chutes team integrated domain specific models so that we can run inference in their verified Trusted Execution Environments (TEEs), with encrypted communication and controlled access to proprietary databases and project records. This architecture is intended to support confidential inference on external infrastructure while preserving organizational control over sensitive research data.
Finding a molecule with promising predicted affinity is only the beginning. Hit-to-Lead optimization involves balancing multiple, often competing objectives, including potency, selectivity, physicochemical properties and synthesizability, before progressing toward experimental activity and ADME validation.
The next step is to integrate the agent into our own Hit-to-Lead development program.
For NOVA, this represents an important step toward translating algorithms developed through open competition into tools for proprietary drug discovery programs, including our own asset-development pipeline and those of external research partners.
CJ 🇺🇸 retweeted
Confidential open-weight models from @chutes_ai sit right next to claude, gpt & gemini on SayGM.
Same key, same code.
Your prompt stays sealed twice: once in our gateway TEE, again in chutes’ attested hardware.
SN28 x SN64 🤝
saygm.com/blog/access-chutes…
Microsoft is essentially firing a third of its company... it spent $41B on Capex last quarter. AI spend is crowding out the labor budgets of major companies. A future where workers compete against expensive GPUs isn't the one I want. AI should be cheap/easy to run + benefit workers.
BREAKING: Microsoft, $MSFT, is offering buyouts to 60,000 workers as part of the company’s efforts to cut costs amid AI adoption, per Bloomberg.
Personnel on select teams in the US, Canada, India, and the UK at a vice president level and below will be eligible.
This reflects ~27% of Microsoft's total workforce.
CJ 🇺🇸 retweeted
We've expanded the Chutes catalog.
As demand for inference has grown, we've added capacity so our users around the world can get the models they need every day, paid for in crypto or fiat. Starting today, alongside our confidential TEE models, Chutes now offers a set of non-TEE models.
First up are the following models, live now on chutes.ai with the same API key and billing you already use.
DeepSeek V4 Pro: chutes.ai/app/chute/chutes-d…
Qwen3.8 Max: chutes.ai/app/chute/chutes-q…
GLM-5.2-Fast: chutes.ai/app/chute/chutes-z…
Qwen3.8-2.4T-A95B: chutes.ai/app/chute/chutes-q…
GLM-5.3: chutes.ai/app/chute/chutes-z…
DeepSeek-V4.1-Flash-1017: chutes.ai/app/chute/chutes-d…
Kimi-K2.7-Code: chutes.ai/app/chute/chutes-m…
More models are coming.
CJ 🇺🇸 retweeted
Here's a very crude simplified chutes roadmap/plan, for near term:
- move the most unprofitable things off of the subnet, including today's announcement of some (not all) additional inference options (not paid via emissions)
- this allows us to scale the things we want and need to avoid capacity issues/better direct incentive - this includes models we use to build data pipelines and relationships to other teams/networks/etc
- add options that are completely revenue positive (hourly pricing/rentals/simple docker based jobs/etc), TEE, private, secure, priced above cost, more extensive version of private chutes we have today basically
- expand the hourly compute offerings to also include CPU only (e.g. to run agents, web browsers, sandboxes, non-GPU work of any kind)
- use some of the incentive for ideally non-datacenter based model training and serving
- build a non-extractable product on chutes from the models we build, via dynamic edge nodes, possibly from the epigenesis work (smaller faster cheaper faster than lora customizations from the sparse sections of pretrained model, participant supplied)
- models and pipeline etc. will be open source, the customization and ensemble thereof will live only in chutes, inc. the continuously evolving set of served variations/RL/etc.
We'd like, to the extent we can, attract participants who want to build and have no obligation to pay datacenters or any particular hardware vendor, to keep the value inside the subnet at all times by attacking sell pressure at the source (DC commitments, chip purchases, etc).
We will continue to have secure TEE inference and GPUs; not elimating anything wholesale, just focusing more on a commodity that doesn't have to flow out to DCs and single hardware vendors.
NVIDIA reportedly has 75% profit margins, revenue for the sake of revenue fueling GPU reselling goes straight to one place, basically. Fine, I guess, but I'd rather work together as individuals with whatever hardware we can bring, to a reasonable degree.
CJ 🇺🇸 retweeted
True? No clue. The point of parallax is to put the network into as many hands as possible in a verifiable trustless way, and for the output (the models) to not require yet another source for the wealthiest companies on earth to monopolize (that is, a model that can run even on a phone and not require "now's" completely unaffordable mega GPU).
The current AI marketplace is fueled by a massive debt market with very few beneficiaries at the world's expense. If you like K shaped economies here it is in plain sight; trickle up resources, consolidating into a few, spreading the gap at ever-increasing speed. Whether this is self serving for Mr. Musk to gather up more resources for pennies on the dollar or a true statement is irrelevant, the outcome is the same.
Will it crash tomorrow, next week, never? No clue, but I would rather not build the foundation on this gamble.
I'm possibly (probably?) doing it wrong, but the goal is and has always been, people first, which means accessibility first. Let's make "good enough" models with no kill switch and no hardware moat. Please and thank you.
If we get superintelligence out into the universe without/at the cost of consciousness to witness it, we have accomplished nothing, and have in fact forgotten the whole point.
Pretty much.
As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.
SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue.
That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.
CJ 🇺🇸 retweeted
We've started rolling out additional models/capacity to the chutes platform, namely Qwen3.8 Max and DeepSeek V4.1 Pro (with more to follow soon, along with additional details). They are currently in beta mode, please test and give us feedback!
Unfortunately, due to the mechanics of these new model offerings and current general GPU/compute market dynamics, subscriptions are no longer feasible. Your subscriptions will automatically sunset/not renew at the end of the current subscription period.
If you have an active subscription it will continue to work normally until the end of the billing period. The two new models listed above are not available with subscriptions and are PayGo only, and they are not currently supported in TEE/confidential compute/E2EE mode.
Please reach out at [email protected] with any questions or concerns.
CJ 🇺🇸 retweeted
Intelligence has quickly become one of the most critical resources on earth, and unfortunately it has already turned into a race for unimaginable wealth and concentrated power. A small handful of companies own the frontier models, the data, and the hardware.
We at Chutes have strong feelings about that. Read our thoughts on a different vision for the world and path forward here: chutes.ai/news/democratizing…
CJ 🇺🇸 retweeted
Railway is compromised, rotate all API keys as soon as possible
Deployments are down so don’t wait to redeploy
Attackers are draining keys in real time, we didn’t lose much but can confirm this is still an active attack
Stay safe friends
CJ 🇺🇸 retweeted
Florian's building super cool stuff, go check it out!
Introducing DecisionModels @decisionmodels_
I've spent the last four weeks benchmarking decision models (I run JevBench).
Today I'm launching the place to actually use the best ones:
decisionmodels.io
➡️Top open models
➡️Up to 4x faster on our own REAX engine
➡️From $0.019 per 1M input tokens.
Expect more reveals soon, and read 👇
CJ 🇺🇸 retweeted
github.com/chutesai/epigenes…
Still experimenting, but seems to check out. Smaller, faster, better? Models that grow with you (or your repo, or proprietary company docs, etc.)
Short term exact recall is the context window.
Context window feeds training continuously, binary optimizer.
This becomes stored knowledge and higher order perceptual understanding over time, not just RAG.
Sparse MSA global attention already selects most important tokens to attend to in longer context, so you can reinforce those parts more.
🤖 Made with AI
There's another magic trick (aka cool math) I believe this parallax enforced sparsity enables by the way...
arxiv.org/abs/1711.05769 <- prune weights, adapt the pruned weights, sharply cuts catastrophic forgetting risk.
If you have a model with the enforced sparsity, 50% of the weights are locked in to particular patterns. What if you mask the other expert weights leaving only those sparse contiguous weights to learn new data?
That's the intuition, perhaps it works, perhaps not. But... if it does, it would make the base model significantly more malleable for new domains/tasks/whatever. SFT/continued pretraining with less risk of catastrophic forgetting and without lots of magic tricks and RL and re-sampling in your own words etc.
CJ 🇺🇸 retweeted
Kappa run is now stopped, not too bad overall! We passed llama 3.2 1b on ARC-C, OBQA, TQA, etc., nearly matched on ARC-E, SciQ, etc. Considering that was 9T tokens and we hit those numbers on the 576B checkpoint, AND we did so around 90% cheaper per token, I'd call it a solid preliminary win!
Issues with the run:
1. In one of the Gated DeltaNet-2 heads, the per-channel decay gate underflowed, causing state wipe, and ourput RMSNorm amplified near-zero outputs ~1000x into fleet wide grad spikes.
2. Very late/slow nodes were so far behind their evidence was discarded, not a huge effect but we should have had better tolerance for stale data and partial updates.
3. We packed samples but didn't do masking/etc. which is sub-optimal for pretraining (not a huge diff but worth fixing anyways).
4. The first and last layers being GDN2 may be suboptimal also, vs. SWA, no clear evidence but might be worth switching up the order a bit for those two.
We have a fix for this GDN2 issue, but while that was still running I also tested EDA (arxiv.org/abs/2606.26560v1) and it actually addresses some of the root cause directly (and seems to have better loss than GDN2 anyways) so we'll probably switch to EDA once all tests come back clean and pass the kappa's instability regime (and waiting until we have highly optimized tilelang kernels for EDA to keep MFU/throughput high).
Onwards and upwards 🚀
CJ 🇺🇸 retweeted
@ExploitSummit was an amazing event! So many brilliant, motivated people.
If you want to hear about the model training stuff we've been working on, and why we're working on it in the first place, here's the talk.
youtube.com/watch?v=4FOxXT66…
If you're concerned about the consolidation of control of AI by a few powerful companies/governments the only alternative is models so small anyone can run one on their laptop, phone or consumer hardware (e.g. like a DGX Spark). This decentralized vision is the future @jon_durbin has set @chutes_ai to building.
Most AI runs in a datacenter, and someone else holds the off switch.
At Exploit Summit (@ExploitSummit ), Jon Durbin showed Parallax running on a phone. CPU only, no GPU. The trained 8B model decoded at 59.6 tokens per second. A 40.75B model ran at 26.9 tokens per second with 12 GB peak process memory.
Where that leads:
• You could use AI in a country that blocks it.
• A doctor could run a model on a tablet in the exam room, and no patient data leaves the building.
• A law firm could keep one on a laptop in a locked office and never worry about prompts leaking to a vendor.
• You could use it on a plane or through an outage.
• No monthly bill and no rate limits. No company deciding tomorrow that your use case violates its policy.
A datacenter has an off switch. A billion personal devices don't.
CJ 🇺🇸 retweeted
"Doing sixty tokens per second on a phone."
At Exploit, @jon_durbin reported an 8B model running on a phone's CPU, without a GPU or NPU. The test used phones rented through Qualcomm Device Cloud, before any optimization work.
CJ 🇺🇸 retweeted
i had no idea what jon durbin’s parallax was or why it mattered until listening to him talk at @ExploitSummit earlier today.
this one talk has been the highlight of the summit thus far.
reconstructing the training pipeline to max joules allowing inference of 50+tps locally from a cell phone is bonkers. puts tokenmaxxing and trainmaxxing to another level.
it’s worth the 1h listen and im looking for excuses to incorporate parallax training to build specialized models to advance the frontier of cyber in an open way
Dashboard is a quick work in progress, but for visibility into a run here ya go!
parallax.chutes.ai/
This is around $11/b tokens, insane actually. MFU also insane. The whole thing, pretty legendary, and inference...
Using a few nodes from @lium_io also!
CJ 🇺🇸 retweeted
Training and inference will run on the edge, by miners, myopically seeking energy wells like they do with Bitcoin mining.
@jon_durbin presenting the future and how humanity gets there.
