AI·Digest
← All digests
evening

AI Digest — July 7, 2026

Quick Notes

  • From Hugging Face to Amazon SageMaker Studio in one click — New deep-link integration lets developers deploy models directly from Hugging Face into pre-configured SageMaker Studio environments with automatic IAM provisioning. See https://huggingface.co/blog/amazon/one-click-to-sagemaker-studio for details on deployment, customization workflows, and integrated GPU quota visibility.

Structured Summaries

Tooling Improvements & Integrations

From Hugging Face to Amazon SageMaker Studio in one click — https://huggingface.co/blog/amazon/one-click-to-sagemaker-studio
Today, we’re excited to announce a deep-link integration between Hugging Face and Amazon SageMaker AI. Developers can now go from model discovery to hands-on experimentation in SageMaker Studio with a single selection. Whether you fine-tune a foundation model (FM) from Amazon SageMaker JumpStart or deploy it to an Amazon SageMaker Inference endpoint, you can now land directly inside the relevant SageMaker Studio workflow. Your selected model is pre-loaded, and the environment is fully configured and ready to go.

Previously, getting started on SageMaker Studio after discovering a model on Hugging Face required navigating multiple steps between opening Amazon SageMaker AI in the AWS Console, creating a domain, configuring IAM permissions, and sometimes requesting GPU quota. For developers who want to iterate quickly, this friction slows down the path from inspiration to experimentation. The integration creates a more direct path from discovery to enterprise deployment.

“At Arcee, we build open models so developers and enterprises can actually own what they run: inspect the weights, post-train on their own data, and deploy on their own terms. This integration takes that promise the last mile. Going from an open model on Hugging Face straight into SageMaker Studio in a single click, then fine-tuning or deploying it inside your own AWS environment with nothing to wire up, is the kind of experience open models have been missing. Open weights you own, running in the cloud you control. That is exactly the combination our customers have been asking for.”

— Mark McQuade, Founder and CEO, Arcee AI

With the launch of a one-click Studio landing experience, choosing Customize on SageMaker AI or Deploy on SageMaker AI on a supported Hugging Face model page takes you directly to the console. SageMaker AI then automatically provisions a new domain with pre-configured permissions in seconds and carries the model context through.

This launch introduces three capabilities that shorten the path from a Hugging Face model to a working SageMaker Studio workflow:

When you browse models on Hugging Face, you’ll now see action buttons alongside supported models that map directly to SageMaker Studio workflows: each entry point preserves the context, meaning you don’t need to search for the model again once inside Studio. New Studio environments created through this flow come with permissions already configured for the full range of SageMaker AI capabilities, including model customization, training jobs, notebook experimentation, and endpoint deployment. A new managed policy, AmazonSageMakerModelCustomizationCoreAccess, is created and attached for you. It provides permissions for serverless model customization jobs using supervised fine-tuning (SFT), direct preference optimization (DPO), reinforcement learning with verifiable rewards (RLVR), and reinforcement learning from AI feedback (RLAIF), with supported deployment to SageMaker AI or Amazon Bedrock endpoints. This alleviates the need to manually create and configure AWS Identity and Access Management (IAM) roles and policies before you can start experimenting. For existing Studio environments, actionable messages with direct links to documentation guide you through adding these permissions. When selecting instance types for deployment or training, the Studio UI now surfaces quota availability directly in the instance selection list. You can immediately see which GPU instance types (G5, G6) are available under your account’s current limits. You don’t need to navigate separately to Service Quotas. If you still need to request a limit increase, you’re redirected directly to the Service Quotas page for the respective instance type. Let’s walk through the experience of customizing or deploying a model starting from Hugging Face: On the Hugging Face model page, click on “Deploy” and select “Amazon SageMaker AI”. If the model is supported, you will see two buttons, “Deploy on SageMaker AI” and “Customize on SageMaker AI”…

Tooling Improvements & Integrations (continued)

sqlite-utils 4.0 release with native schema migrations — https://simonwillison.net/2026/Jul/7/sqlite-utils-4/#atom-everything
This morning I released sqlite-utils 4.0, the 124th release of that project and the first major version bump since 3.0 in November 2020. In addition to some small but significant breaking changes (described in this upgrade guide), this version introduces three major features: database migrations, nested transactions (via a new db.atomic() method), and support for compound foreign keys.

Schema migrations define a sequence of changes to be made to a SQLite database, plus a mechanism for tracking which migrations have been applied and applying any that are found to be pending. Migrations are defined in Python files using the sqlite-utils Python library, which includes a powerful table.transform() method providing enhanced alter table capabilities that are not supported by SQLite’s ALTER TABLE statement (table.transform() implements the pattern recommended by the SQLite documentation—create a new temporary table with the new schema, copy across the data, then drop the old table and rename the temporary one in its place).

Here’s an example migration file which creates a table called creatures, adds an additional column to it in a second step, then changes the types of two of the columns in a third: Save that as migrations.py and run it against a fresh database like this. Then if you check the schema of that database, The _sqlite_migrations table is used to keep track of which migration functions have been run. The creatures table above is the schema after all three migrations have been applied.

To see a list of migrations, both pending and applied, run this: If you don’t specify a migrations file, the sqlite-utils migrate data.db command will scan the current directory and its subdirectories for files called migrations.py and apply any Migrations() instances it finds in them. You can also execute migrations from Python code using the migrations.apply(db) method, which is useful for building tools that manage their own database schemas over multiple versions. My own LLM tool has been using a version of this pattern for several years now, as shown in llm/embeddings_migrations.py.

My favorite implementation of this pattern remains Django’s Migrations, developed by Andrew Godwin based on his earlier project South. Fun fact: Andrew, Russ Keith-Magee, and I presented our competing approaches to schema migrations for Django on the Schema Evolution panel at the very first DjangoCon back in 2008! My attempt was called dmigrations, developed with a team at Global Radio in London.

Django’s migrations can be automatically generated from model definitions and include the ability to roll back to a previous version. The sqlite-utils approach is deliberately simpler: unlike Django, sqlite-utils encourages programmatic table creation rather than a model definition ORM, so there isn’t anything we can use to automatically generate migrations. I decided to skip rollback, since in my experience it’s a feature that is rarely used. With a SQLite project, an easy way to achieve rollback is to create a copy of your database file before you apply the migrations!

The design of sqlite-utils migrations is three years old now—I had originally released it as a separate package called sqlite-migrate, which never quite graduated beyond a beta release. I’ve used that package in enough places now that I’m confident in the design, so I’ve decided to promote it to a feature of sqlite-utils to make it available by default to all of the other tools in the growing sqlite-utils/Datasette/LLM ecosystem.

I made one last release of sqlite-migrate, which switches it to depend on sqlite-utils>=4 and replaces the init.py file with the following: Any existing project that depends on sqlite-migrate should continue to work without alterations. Here are the release notes for this version, with some inline annotations: The 4.0 release includes some minor backwards-incompatible fixes (hence the major version number bump) and introduces three major new features…

Tooling Improvements & Integrations (continued)

sqlite-migrate 0.2 final release — https://simonwillison.net/2026/Jul/7/sqlite-migrate/#atom-everything
The version that retires the library, instead implementing a compatibility shim against the new sqlite-utils 4.0 dependency…

Developer Tools & Experiments

github-code Web Component — https://simonwillison.net/2026/Jul/7/github-code-component/#atom-everything
An experimental Web Component built using GPT-5.5 and the following prompt: let’s build a Web Component for embedding code from GitHub

It takes URLs like that, converts them to https://raw.githubusercontent.com/simonw/sqlite-ast/437c759129154f05296324a7f82aa1246340dd14/sqlite_ast/parser.py, then uses fetch() to fetch them and displays the specified range of lines - with line numbers, no syntax highlighting though

Show me a preview web browser so I can see your work… Here’s what it looks like embedded on this page: This is a beat by Simon Willison, posted on 7th July 2026. Sponsor me for $10/month and get a curated email digest of the month’s most important LLM developments.

Research & Industry Analysis

The Big Ways AI Just Changed — https://www.youtube.com/watch?v=WUeMqk_ABmY
Today on the AI Daily Brief, why June was the most significant month in AI in years. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Well, friends, it is July 4th weekend. Most of my American listeners, at least, are washing summertime lakes, fireworks, and patriotic feelings on the 250th anniversary of this country. And yet, over here in AI, we are shifting from one month to another. It is a little poetic that at the very, very end of the month, Fable 5 got back just in time for us to get back to building in July. And yet, before we do so, it is worth spending a moment, I believe, looking back at the last month, which I would argue is one of the most significant in the post-ChatGPT history of AI. By the way, for those of you wondering, this website companion experience was actually created not with Fable, but with Codex and GPT-5. Before we get into June, let’s actually go back to May. The historian in me thinks that these two months kind of make a matched pair, telling the same story but from different angles. So, the story of May was all about the shift from the AI subsidy era to the token scarcity era. Even before May, we had started to see providers shift away from their seat-based subscription models and move towards more usage-based models. This was, of course, the inevitable consequence of shifting from pre-agentic to agentic workloads, which consume just an absolutely massive amount more of intelligence than the type of queries that we were running back in ‘24 and ‘25. May was also when we started to see the chickens coming home to roost when it came to enterprises that had run out to start token maximizing. Uber had been in the news for a couple months asit burned through its AI budget in the first 4 months of the year. We got more and more reports of companies turning off their token leaderboards, and all in all, May felt like the beginning of a shift to a new paradigm…

Research & Industry Analysis (continued)

AI Focused Companies Are Hiring More — https://www.youtube.com/watch?v=PnEWo_AKzIU
Today on the AI Daily Brief, the latest numbers on AI and jobs. Before that, in the headlines, is OpenAI about to give 5% of the company to the US government? The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI? Welcome back to the AI Daily Brief headlines edition. All the daily AI news you need in around 5 minutes. Well, all those conversations about AI companies getting more tied up with the government are getting a lot more real. OpenAI has proposed handing over a 5% stake in the company to the US government. Now, obviously, there has been chatter about this for some time now. And honestly, if you go way way back in Sam Alman interviews from 3 or 4 years ago, it was clear at the very beginning of the project that he thought this sort of pursuit of AI was something that inherently should have government involvement from the very beginning. I can’t find the exact interview, but I remember thinking that he almost seemed surprised that the US government wasn’t interested in getting involved back then. Now coming back to the deal now… This potential 5% stake would be contributed to a sovereign wealth fund structured in a similar way to the Alaskan permanent fund whichcollects oil and mining revenue for the benefit of citizens. At current valuations, the stake would be worth around $42 billion. And it’s unclear whether the administration would be purchasing the stake or whether it would be a gift to the American people… The FT pointed to what many people assume is a pretty clear quid pro quo, writing, giving the government an ownership state could help secure good relations with the administration and would mark an attempt…

AI in Creative Fields

Regenerative beekeeping with Claude — https://www.youtube.com/watch?v=CuzipWWYvFo
I’ve been doing regenerative beekeeping for about 10 years. There are parts of this job that are instinct, but there’s also a huge amount of information, and [music] Claude helps me hold both at the same time. >> Oh my god, I’m so happy. << So, this is swarming. This is how bees naturally reproduce. It’s like the hive is giving birth to itself. Regenerative beekeeping is really looking at how bees live in the wild. When they leave, they’ll cluster somewhere like a branch, and that’s where I would get a call. >> Yes, I do. Right in the front yardin a ginkgo tree… Beekeeping, in my opinion, is mostly observation. And the more you sit with them, the more you get to know them. You get to know what’s normal for them. They’re just looking for a new home. >> I know… Every single situation, every single scenario is completely different.** Claude has been helping me consolidateall of this data that I’ve been collecting of commonly asked questions and help me put it together in a one-page fact sheetthatI [music] can easily hand off to a client… << Where is the queen? >> She’s somewhere in there. It really forces you to [music]use your intuition and also to have a relationship because there’s no blanket solution. So, I really have to sit with it and sit with the individualhive and the individual scenario and ask myself like,“What makes the most sense right now?” There’s only so muchyou can control. I think it’s more about stepping into a partnership with nature…**The deeper you look, themore interconnections [music]you see between bees and everything else. I feel like I’m just scratchingthe surface…

🔗 View this digest on the web: https://ai-digest-b7u.pages.dev/digests/2026-07-07-evening/

#ai#digest