Companies are loving our eBOL Specific Scanner

I mean who would not? It turns paperwork into structured digital data in just seconds

Output Speed
- Output tokens per second · Higher is better
- Comparison of first-party API performance

The Subject: AI writing in Linkedin Posts

The Verdict: Lazy, cheap, obvious.

The Breakdown:
You can spot them in two seconds flat. Every post starts with the exact same fake enthusiasm, five paragraphs of generic bullet points, and a cheesy summary that tells you nothing. People post them because they think it makes them look smart and consistent. In reality, it just tells everyone in your network that you do not care enough to share an actual human thought.

Bending Spoons is Buying Airtable

At its peak, Airtable was known as a SaaS company that was doing “everything right.” In 2021, it raised capital at a post-money valuation of $11.735 billion.

This morning the firm announced it had agreed to be acquired by Bending Spoons for an enterprise value of $1.285 billion.

What Airtable actually sells

Airtable is a relational database with a top hat. It makes that database as friendly and accessible as a spreadsheet. (This sounds simple, but is actually quite a hard product to pull off.) To upsell it beyond hobbyists, the company would bundle in the usual stuff like permissions, integrations, and internal application building.

Tools like Lovable, Replit, Claude Code, and ChatGPT can increasingly generate everything you need besides the database itself. For a small dataset, a company can use Notion or something inexpensive. If the application gets too big, your AI agent can connect you to Snowflake, Databricks, or a host of other options.

Monthly spending on data center construction in the United States

I tried Kimi first time yesterday - Unlike ChatGPT or Claude, it was genuinely neutral and does not try to impress you.

Has anyone tried? If yes, what's your take?

Somewhere right now, a dispatcher is squinting at a driver's handwriting trying to read a BOL. Somewhere else, a driver's waiting on him to figure it out. We fixed that

#logistics

Running Kimi K3 locally is easy.
You just need:

- 1.5 TB of GPU memory
- roughly 8H100s
- A small power substation

Olix raises $312M at $3.3B valuation from Netflix’s Reed Hastings, Arm, to build Nvidia rival

Olix says it treats a data centre running AI inference like a factory that produces tokens. Each token needs hundreds of operations and different hardware.

Most companies use a single general-purpose chip for everything, but Olix’s X-1 platform distributes the model across several specialised chips, each handling a specific task. The system uses a flexible compute fabric instead of locking in a model’s design.

The chips use an optical interconnect, sending data via light rather than copper. Olix says this lowers latency and saves energy. The first chip, DX-1, is a decode accelerator for the stage when a model creates its output.

For models with 100 billion parameters, Olix claims that DX-1 can deliver over 10,000 tokens per second per user and use less power than general-purpose chips for large batches. The design is meant to scale to models with 10 trillion parameters or more.

Olix has over 140 employees in London, Bristol, Toronto, Austin, and San Francisco. The new funding will help launch DX-1 to its first customers in the second half of 2027, and support more platform development and manufacturing to increase chip production.