
Doug Barton
3.5K posts

Doug Barton
@bartond
Director, University of Wisconsin E-Business Consortium; EmTech exec where my passion is helping people and organizations achieve their potential w/innovation.



vibe slopers can't into tmux




(5/8) This leaves us with a mix of excitement and humility. 1. AI is especially good at connecting distant fields of research. Here it finds bridge between algebraic number theory and discrete geometry. We hope to see more bridges built by AI connecting different fields.



Chamath just delivered the clearest diagnosis of what is happening to enterprise software and the OpenAI Deployment Company is the most damning piece of evidence he could have picked. "The low end of the market is basically finished. There is no safe space." 90% of public SaaS stocks are down 30-80% from their 52 week highs, the median software stock is now negative over the last 3-6 months. Goldman Sachs reported that software forward P/E multiples fell from 35x to 20x, the lowest absolute level since 2014 and the smallest premium to the S&P 500 since 2010. The low end died first and fastest, because AI replaced it most directly. The small business tools, the lightweight project managers, the single function SaaS products that charged $49 a month per seat, those are being replaced by AI agents that do the same work as a workflow, not a product. You do not buy an AI powered tool, you describe what you need and it builds it and the seat based model that created the SaaS industry simply does not apply to that transaction. But Chamath's more interesting argument is about the high end and the tell he points to is perfect. OpenAI just raised $4 billion from 19 investors including TPG, Brookfield, Bain, and McKinsey to launch a consulting company and guaranteed those investors a 17.5% annual return to do it. On $4 billion in committed capital, that is roughly $700 million per year in guaranteed payouts, owed by a company that is projected to lose $14 billion in 2026. The goal of this venture is to compete directly with Deloitte, PwC, Ernst & Young, Andersen, and Cognizant. Think about what that structure reveals. OpenAI lost half of its enterprise LLM API market share from 50% to 25% between late 2023 and mid-2025, with Anthropic now leading at 32%. Its response was not to build a better model but rather to raise $4 billion, offer guaranteed PE-tier returns and hire embedded engineers to physically sit inside client organizations and make AI actually work in production. The reason, as Chamath identified, is that the high end of the market is not easy. "It's not like boop boop boop, put in a prompt and beep bap boop, it all works," he said and the data confirms exactly that. 88% of organizations running AI agents reported a security incident in the past year, 42% of C-suite executives say AI adoption is creating internal organizational conflict. The average enterprise AI consulting implementation costs $228,000 in year one versus $77,000 for platform-based approaches and most still stall before reaching production. Anthropic immediately matched OpenAI with a competing $1.5 billion consulting venture backed by Blackstone, Goldman Sachs, and Hellman & Friedman bringing the combined spend by the two leading AI labs on human powered enterprise deployment to $5.5 billion in a single month Chamath's read is that the high end, the large enterprise platforms like Salesforce with proprietary data flywheels, Palantir with its FDE model already proven at scale, Oracle with vertical specific data moats will survive and consolidate. The mid-market point solutions, the single function tools, the lightweight enterprise apps without defensible data assets, those are on the conveyor belt. The AI industry is not just disrupting the companies that use software but rather disrupting the companies that sell it.


Major new feature for @NotebookLM power users: in the tradition of Mind Maps, Notebook can now auto-label your sources, making it much easier to manage notebooks with many sources. I’ve been using it for weeks and it is amazingly versatile with big notebooks. Details below. Here’s how it works. If you have more than 5 sources in a Notebook, you’ll see a new “auto-label” button above the source list on the left side. Click on that and Notebook will review the content of all your sources and organize them into high-level categories. Each source can have multiple labels if there is overlap in the subject matter. Once the labels have been applied, you’ll see a new tidy view of your sources where you see only the top level categories, but you can easily expand to see all the sources associated with each label. Click the three dot menu next to each to rename or delete the label. (Sources won’t be deleted.) Or add emojis to visually differentiate between labels. You can click the three dot menu next to each source to assign different labels to the source. Having that organized label view in the source panel makes it much easier to find a specific source you’re looking for, but that’s just the beginning. You can also focus the AI on specific categories using the selection buttons on the right hand of the source panel. Select one category and all the responses in chat will be grounded exclusively in the sources assigned to that label. This can be helpful if you’re worried about the AI getting distracted by information in other categories, and it can speed up your chat response times because there are fewer sources to load into the context. Selecting by label is also super helpful for generating studio artifacts. If you want a podcast focused only the sources about the civil war in your American History notebook, just select that label and click the audio overview button in Studio. Label view also greatly enhances Fast and Deep Research in a notebook with many existing sources. In the past, if one of the research agents added a batch of sources (up to 40 or 50 with Deep Research) all the sources would be scattered through your source panel alphabetically with no way to tell which ones were the new additions. But now, if all your pre-existing sources are neatly filed away in the appropriate labels, when you pull down new research sources they all appear in alphabetical list below the label categories. That makes it easy to review those new sources to see which ones you really want to keep, and you can manually select them (and de-select all the labels) to explore the new information you’ve just added to your notebook. Let’s say you want to add new information specifically about the Battle Of Gettysburg to your American History notebook—run a Fast Research query, import ten new sources, select those new unlabelled sources and hit the Slide Deck button to do a focused review of the history of Gettysburg. Once you’ve explored those new sources, you can always hit the original auto-label button in the top left and choose “Reorganize unlabeled sources.” Notebook will automatically assign the appropriate labels to the new arrivals. If you want to switch back to the full alphabetical list of sources, just choose “Return to list view” to return to the traditional source panel layout. Notebook will remember your labels so it's easy to switch back and forth between the two views. The feature should be rolling out to all users over the next few days. Enjoy!














Moral of the story is not to use LLM-as-judge without determining human correlation or tuning to maximize that








