Mira Murati's Thinking Machines released Inkling, a 975 billion parameter sliding-window mixture-of-experts model, as open weights on HuggingFace under Apache 2.0; this video separates the genuine win (the first credible American frontier-scale open-weights model) from the catch (roughly 300GB of RAM even quantized, and rental pricing at $45 per million output tokens).
The Stack11 minTranscript found
Quick learning frame
Read this before watching.
Creative automation uses agents to accelerate production while keeping human taste in story, pacing, selection, and critique.
New playlist item from The Stack; queued for transcript-backed review, topic mapping, and a practical learning artifact.
Skill you build: The ability to read an open-weights release honestly, separating license freedom from hardware feasibility, vendor-reported benchmarks, and the hosting and fine-tuning business the free weights are really advertising.
Watch for the shift from claim to mechanism. The learning value is the point where the transcript reveals a repeatable action, tool boundary, context move, review habit, or artifact.
Concept diagram
Where this video fits.
01Brief
02Source
03Generation
04Selection
05Edit
06Taste Review
Deep lesson
Turn this video into working knowledge.
1,678 cleaned transcript words reviewed across 534 timed caption segments.
Thesis
The Free 975B AI Model You Can Own (Inkling AI) teaches a practical creative automation move: Mira Murati's Thinking Machines released Inkling, a 975 billion parameter sliding-window mixture-of-experts model, as open weights on HuggingFace under Apache 2.0; this video separates the genuine win (the first credible American frontier-scale open-weights model) from the catch (roughly 300GB of RAM even quantized, and rental pricing at $45 per million output tokens).
The goal is not to remember the video. The goal is to extract the operating principle, tie it to timestamped evidence, test how far the claim transfers, and make something reusable.
0:00
Sparse trillion scale
“Mera Maratti, OpenAI's former CTO, just dropped a 975 billion parameter model called Inkling. And instead of locking it behind an API, her new lab uploaded the open weights directly to HuggingFace, totally free under an Apache 2.0...”
Inkling is 975 billion total parameters trained on 45 trillion tokens across text, images, audio and video with up to a 1 million token context, but it uses a sliding window mixture-of-experts design where only six of 256 routed experts fire per token, so roughly 41 billion parameters are active at inference; it also ships a controllable thinking-effort dial from 0.00 to 0.99 that trades compute for accuracy per request, plus a preview Inkling Small with 12 billion active parameters. Work out the active-to-total parameter ratio (about 41B of 975B) and compare it against a dense model you already run to see what actually has to sit in memory.
5:37
Own it to change it
“model yourself. On paper, the open- source ecosystem was completely ready for this thing to drop. You had immediate day zero support in llama.cpp CPP for local inference, which means the community was already wiring this up before...”
The point of holding the weights is weight-level control a closed API contractually forbids: you can inspect what is inside, audit exact behavior for enterprise compliance, and permanently alter the model on your proprietary data using standard tools like Unsloth or the lab's companion fine-tuning platform Tinker, and the launch demo had Inkling write its own fine-tuning job, run it on Tinker and evaluate the results. The trade is giving up a few points of general capability for a base you can specialize. Name one proprietary dataset you would fine-tune on and write down exactly what a closed API would not let you do with it.
7:17
Free license, rented meter
“audit its exact behavior for enterprise compliance, and most importantly, permanently alter it to fit your exact workflow. Thinking Machines designed the release for this exact loop, shipping the model alongside their companion fine-tuning platform, Tinker, which means...”
Only two third-party providers host Inkling, Together AI and OpenRouter, plus the lab's own Tinker, and renting costs about $1 per million input tokens and $45 per million output, roughly four times comparable open-weight models on the same page; self-hosting instead means around 300GB of RAM for heavily quantized community builds or terabytes of GPU VRAM for the full BF16 checkpoint, so for most users the open model is still a rented model at closed-model rates. Price your own monthly token volume at $1 in and $45 out, then compare it against renting the GPU capacity a 300GB quantized build would need.
01
Brief
Start with this video's job: Mira Murati's Thinking Machines released Inkling, a 975 billion parameter sliding-window mixture-of-experts model, as open weights on HuggingFace under Apache 2.0; this video separates the genuine win (the first credible American frontier-scale open-weights model) from the catch (roughly 300GB of RAM even quantized, and rental pricing at $45 per million output tokens). Treat "Brief" as the outcome you are trying to make visible, not a topic label. Anchor it to 0:00, where the video says: “Mera Maratti, OpenAI's former CTO, just dropped a 975 billion parameter model called Inkling. And instead of locking it behind an API, her new lab uploaded the open weights directly to HuggingFace, totally free under an Apache 2.0...”
02
Source
Use "Source" to locate the part of the creative automation workflow the video is demonstrating. Ask what changes in your real setup if this claim is true. Anchor it to 5:37, where the video says: “model yourself. On paper, the open- source ecosystem was completely ready for this thing to drop. You had immediate day zero support in llama.cpp CPP for local inference, which means the community was already wiring this up before...”
03
Generation
Turn "Generation" into the reusable artifact for this lesson: A creative workflow board with critique criteria and review checkpoints. This is where watching becomes something you can inspect and reuse.
04
Selection
Use "Selection" as the application surface. Decide whether the idea touches a browser flow, a local file, a model choice, a source document, a UI, or a review step.
05
Edit
Use "Edit" to prove the lesson. The evidence should connect back to the video title, transcript anchors, and a concrete output, not a generic best-practice claim.
06
Taste Review
Use "Taste Review" to carry the idea forward: save the prompt, checklist, diagram, or operating rule that would make the next agent run better.
Example
Source-backed work packet
Convert the video into a scoped task that includes the transcript claim, target workflow, acceptance criteria, and proof. The output should be a creative workflow board with critique criteria and review checkpoints..
Example
Claim vs. demo brief
Separate what the speaker claims, what the demo actually proves, and what still needs outside verification before you adopt the workflow.
Example
Teach-back module
Transform the lesson into a definition, a mechanism diagram, one misconception, one practice exercise, and a check-for-understanding question.
Do not learn it wrong
Treating the title as the lesson without checking what the transcript actually says.
Letting the prompt drift into generic advice that could apply to any video in the playlist.
Copying the tool setup without identifying the operating principle that transfers to your own stack.
Skipping the artifact, which means the learning never becomes operational or inspectable.
Do not count this as learned until these are true.
01
State the transcript-backed claim in your own words: Mira Murati's Thinking Machines released Inkling, a 975 billion parameter sliding-window mixture-of-experts model, as open weights on HuggingFace under Apache 2.0; this video separates the genuine win (the first credible American frontier-scale open-weights model) from the catch (roughly 300GB of RAM even quantized, and rental pricing at $45 per million output tokens).
02
Explain the practical stakes without hype: New playlist item from The Stack; queued for transcript-backed review, topic mapping, and a practical learning artifact.
03
Map the idea onto the Brief -> Source -> Generation -> Selection -> Edit -> Taste Review sequence and name the weakest link.
04
Produce the artifact and include the evidence that proves it: A creative workflow board with critique criteria and review checkpoints.
Put it into practice
Give this grounded prompt to Codex or Claude after watching.
You are helping me turn one specific YouTube video into real, durable learning.
Source video:
- Title: The Free 975B AI Model You Can Own (Inkling AI)
- URL: https://www.youtube.com/watch?v=E8eLES1NCa4
- Topic: Creative Automation
- My current learning frame: Take one workload you currently send to a closed API, estimate its monthly cost at Inkling's $1 and $45 per million token rates, compare that against self-hosting a quantized build needing roughly 300GB of RAM, and decide which form of ownership you are actually buying.
- Why this matters: New playlist item from The Stack; queued for transcript-backed review, topic mapping, and a practical learning artifact.
Transcript anchors from this exact video:
- 0:00 / Evidence 1: "Mera Maratti, OpenAI's former CTO, just dropped a 975 billion parameter model called Inkling. And instead of locking it behind an API, her new lab uploaded the open weights directly to HuggingFace, totally free under an Apache 2.0..."
- 2:37 / Evidence 2: "on the artificial analysis intelligence index, landing exactly three points above Neatron 3 Ultra. That makes it the new leading American open weights model at release, which is a massive deal if you care about where your foundation..."
- 5:37 / Evidence 3: "model yourself. On paper, the open- source ecosystem was completely ready for this thing to drop. You had immediate day zero support in llama.cpp CPP for local inference, which means the community was already wiring this up before..."
- 7:17 / Evidence 4: "audit its exact behavior for enterprise compliance, and most importantly, permanently alter it to fit your exact workflow. Thinking Machines designed the release for this exact loop, shipping the model alongside their companion fine-tuning platform, Tinker, which means..."
- 9:34 / Evidence 5: "regulated workflow, the answer is genuinely yes. But for the rest of us building things at home, the honest translation is that someone can own it. And that still matters because a strong American base model existing at..."
Your task:
1. Use the transcript anchors above as the primary source packet. If you add outside context, label it clearly as outside context and keep it secondary.
2. Create a source-check table with columns: timestamp, claim, what the demo proves, confidence, and what still needs verification.
3. Extract the actual teachable claims from the video. Do not invent claims that are not supported by the title, lesson frame, or transcript anchors.
4. Build a reusable learning artifact: A creative workflow board with critique criteria and review checkpoints.
5. Include:
- a plain-English definition of the core idea
- a diagram or structured model using this sequence: Brief -> Source -> Generation -> Selection -> Edit -> Taste Review
- 3 concrete examples that apply the video idea to real agentic work
- 2 failure modes the video helps prevent
- a checklist I can use the next time I run Codex or Claude
- one practical exercise with a clear done signal
6. Add a "learning transfer" section: what changes in my workflow tomorrow if I actually learned this?
7. Add a "source check" section that cites which transcript anchor supports each major takeaway.
Quality bar:
- Make this specific to "The Free 975B AI Model You Can Own (Inkling AI)", not a generic Creative Automation essay.
- Prefer operational examples, failure modes, and reusable artifacts over broad definitions.
- Call out uncertainty instead of smoothing over weak evidence.
- If evidence is weak, say what transcript segment or timestamp needs review instead of guessing.
- Finish with a concise artifact I could paste into my learning app.
Misconceptions
What to stop believing.
Creative AI removes the need for taste.
It increases the need for taste because output volume explodes.
The best prompt is enough.
References, critique, iteration, and post-production matter just as much.
Practice studio
Learning only counts when you make something.
01
Transcript evidence map
Separate what the video actually says from what you already believe about the topic.
3 source-backed takeaways with timestamps, confidence, and a transfer note.02
One useful artifact
Apply the video to a real workflow and produce a creative workflow board with critique criteria and review checkpoints..
A reusable artifact with a done signal and one verification step.03
Teach-back card
Explain the lesson to someone who has not watched the video yet.
A 90-second explanation, one diagram, one example, and one misconception to avoid.
Recall check
Answer first, then reveal — without rewatching.
How many of Inkling's 256 routed experts fire per token, and how many parameters end up active?
What does holding the weights let you do that a closed API contractually prohibits?
Why does the video describe the free license as a beautifully executed trap?
Source shelf
Use the video as a doorway, then verify with primary sources.
