Michael Bleigh

@mbleigh.dev

Building the servers for serverless at Firebase. Web platform aficionado.

$import { genkit } from "genkit"; import { googleAI } from "@genkit-ai/google-genai"; const plant = fs.readFileSync('palm_tree.png', { encoding: 'base64' }); const room = fs.readFileSync('my_room.png', { encoding: 'base64' }); const ai = genkit({googleAI()}); const { media } = await ai.generate({ model: googleAI.model('gemini-2.5-flash-image-preview'), prompt: [ { text: 'add the plant to my room' }, { media: {url: `data:image/png;base64,${plant}`} }, { media: {url: `data:image/png;base64,${room}`} }, ], config: {responseModalities: ['TEXT', 'IMAGE']}, }); media.url; // "data:image/png;base64,..."$

A man in a hot dog costume saying "We're all trying to find the guy who did this."

We need to rethink our listing process at
@coinbase
given there are ~1m tokens a week being created now, and growing. High quality problem to have, but evaluating each one by one is no longer feasible. And regulators need to understand that applying for approval for each one is totally infeasible at this point as well (they can't do 1m a week). It needs to move from a an allow list to a block list, and utilize customer reviews/automated scans of onchain data etc to help customers sift through.

That and we'll continue integrating native DEX support more deeply. Customers shouldn't need to know or care whether the trade is happening on a DEX or CEX.

$type Product @table { name: String!, listPrice: Int! } type Order @table { purchaseTime: Timestamp! orderTotal: Int } type OrderItem @table { order: Order! product: Product! quantity: Int purchasePrice: Int # fixed at checkout time } type DailyOrderReport @view(sql: """ SELECT DATE_TRUNC('day', o.purchase_time) as order_date, COUNT(DISTINCT o.id) as total_orders, SUM(oi.quantity) as total_items, ROUND(AVG(o.order_total)) as avg_order_total, SUM(o.order_total) as total_revenue FROM order o LEFT JOIN order_item oi ON oi.order_id = o.id GROUP BY DATE_TRUNC('day', o.purchase_time) """) { orderDate: Date @unique totalOrders: Int totalItems: Int avgOrderTotal: Int totalRevenue: Int } # now we can create queries to expose easily chartable information query DailyOrders( $startDate: Date, $endDate: Date ) @auth(expr: "auth.token.admin == true") { dailyOrderReports(where: {orderDate: {gte: $startDate, lte: $endDate}}) { orderDate, totalOrders, totalItems, avgOrderTotal, totalRevenue } }$

$type Movie @table { title: String! releaseDate: Date! } type Review @table { movie: Movie! rating: Int } type ReviewStats @view(sql: """ SELECT movie_id, COUNT(*) as count, ROUND(AVG(rating), 2) as avg_rating """) { movie: Movie @unique count: Int avgRating: Float } # now I can add review stats to any query about movies… query LatestMovies @auth(level: PUBLIC) { movies(orderBy: [{releaseDate: DESC}], limit: 10) { id, title reviewStats { count, avgRating } } }$

type Article @table { title: String! body: String! tags: [String] }

type Tag @view(sql: """
SELECT tag, COUNT(*) as count
FROM article, unnest(tags) as tag
GROUP BY tag
""") {
tag: String @unique
count: Int
}

# now we can query it like any other type…

query TopTags(
# if supplied, find tags with prefix e.g. "news-". if omitted, matches all tags
$prefix: String
) @auth(level: PUBLIC) {
tags(
where: {tag: {startsWith: $prefix}},
orderBy: [{count: DESC}], limit: 5
) { tag, count }
}

$type MyView @view(sql: """ -- you can write an arbitrary SELECT statement as a view SELECT field1, user_id, SUM(field2) as total FROM my_table GROUP BY field1, user_id """) { # then map it to GraphQL fields field1: Int # or even define relations to other types user: User total: Int }$