You built a skill for your technology. API references, authentication flows, SDK patterns, error handling, version info, all packed into one skill. The agent calls it, gets all that context, and generates code. The kicker? You’ve just wasted a lot of tokens.

It already knows

Models have ingested your documentation, your Stack Overflow answers, your GitHub repos, your blog posts. The default imports, the standard auth flow, the common CRUD operations: the model already has all of that baked in.

When your skill repeats what the model already knows, you’re not helping, you’re adding weight. Every token your skill returns occupies space in a finite context window, and those tokens aren’t neutral. They push out the stuff the model doesn’t know: workspace files, conversation history, or output from other tools.

Most skills suffer from the same problem. Stuffed with thousands of tokens of documentation, called by the agent, payload returned, and outcomes don’t improve. Sometimes they get worse, because the skill is doing work the model didn’t need help with.

How do you know what the model knows?

You don’t, unless you measure. And most folks skip this step entirely. They go straight from “we have a technology” to “we need a skill” without checking what the model does on its own.

What if the model already generates correct code for your API 90% of the time? Then you need a lightweight skill that covers the remaining 10%: the auth quirk that trips people up, the breaking change that’s too recent for training data, the configuration pattern that looks like nothing else on the web. You can’t know which 10% to target if you haven’t measured the baseline.

Measure first, build second

Start by running your scenarios without the skill. Same model, same harness, same prompts. See what the agent gets right and what it gets wrong, because that’s your baseline.

If the model handles CRUD correctly, don’t put CRUD examples in your skill. If auth flows work out of the box, don’t include your auth guide. If it picks the right SDK version, don’t waste tokens telling it which version to use.

What’s left after you subtract the baseline? The patterns the model gets wrong or doesn’t know about at all. That’s your skill’s scope. Nothing more.

Every unnecessary token is drag

Context windows have a fixed budget. A skill that returns 3,000 tokens of documentation the model already knows is burning context that could hold the developer’s workspace files, conversation thread, or output from another tool the model needs.

It gets worse when skills compose. Your developers have other skills installed, and each one claims tokens just by being present. Your oversized skill isn’t just dragging its own scenarios, it’s eating into the budget other skills need. You’re not just hurting your outcomes, you’re contributing to everyone else’s drag.

The lean skill

1. Define scenarios: the tasks developers actually ask agents to do with your technology.
2. Run them without your skill, and score the outcomes.
3. Identify where the model fails: those failures are your scope.
4. Build a skill that addresses only the gaps.
5. Measure again. Confirm you’re producing lift, not drag. Also pay attention to the token count: lift at 3x the tokens is a net loss.

Do this and you’ll end up with skills a fraction of their original size that produce measurably better results. In many cases, models don’t need a textbook. They needed a cheat sheet.

The post Stop overloading your skills appeared first on Microsoft for Developers.

How to build a production-grade HTTP server without async/await or coroutines Editor’s note: This is a guest post by Peter Mbanugo (originally published on Peter’s blog). Peter will be speaking at P99 CONF 2026. His topic: Multi-Core Without the Trilemma: Escaping Async/Await, Mutexes, and GC. Register now — the conference is free and virtual. In previous articles, we explored why async/

Source

(Kimberly (she/her) took the express train down the fountain pen/stationery rabbit hole and doesn't want to be rescued. She can be found on Instagram @allthehobbies because there really are many, many hobbies!.)

I was excited to pick up the Piper Trading Company inks at the Philly Pen Show earlier this year and I’m finally getting around to reviewing them. Thank you to Scott Franklin of Franklin-Christoph and Piper Trading Company for providing these inks for review.

Piper Trading Company (PTC) is a wholly-owned subsidiary of Franklin-Christoph and had its first show debut at the St. Louis show last year. Full disclosure: I have worked behind the Franklin-Christoph tables for many years and also for PTC’s first show last year and will again be doing so at this year’s St. Louis show. As with all of my reviews, my views are my own and I was not compensated by F-C or PTC for this review.

This is PTC’s first foray into inks; there are 4 inks so far: Fickle Pickle, Firefox, Kissin’ Cousin, Mysteria Wisteria. They are sold in 30ml/1ounce bottles and cost $10 per bottle. The four bottle set is currently unavailable.

As usual, all swatches were done on Col-O-Ring and Wearingeul Instant Film Color swatch cards using a Kakimori steel dip nib and writing samples were done primarily with a Lamy Vista with a steel Medium nib and a TWSBI Go with a Medium nib. The notebooks used for writing samples are from an Endless Recorder and Odyssey Notebooks, both with 68 gsm Tomoe River paper. Dry times for the Vista are shown with “(V)” and the Go will be shown below that with a “(T)”. Dry times may be a bit slower on 52gsm TR or wetter nibs or faster on more absorbent papers like Rhodia, copy paper, Cosmo Air Light, or with drier or finer nibs.

Fickle Pickle is a medium olive green that is neither too light, nor too dark. It is a nice shading ink that is almost a chromashader with hints of brown in heavier swatches. Flow was average.

Firefox is a slightly orange-leaning red that can show up as an almost pastel coral pink with drier nibs. Flow felt slightly dry in the Vista but not in the Go.

Kissin’ Cousin is a bright magenta that also has a bit of lightness/softness to it. It had average flow despite it being a shader (which can sometimes feel dry). Dry time was pretty fast too.

Mysteria Wisteria is a vivid blue-purple with good flow and very fast dry times.

All of the inks had average flow, with the exception of Firefox in the Vista, which I’m going to chalk up to the pen, since it was fine in the Go. Can’t decide between Fickle Pickle and Mysteria Wisteria as my favorites of the bunch. At $10 for a 30ml bottle, they are a good value especially if you don’t already have similar colors. You can buy Piper Trading Company inks online or at the St Louis Pen Show next weekend. Please stop by and say hi if you’re going to the show!

(Disclaimer: Thank you to Scott Franklin for providing these inks for review. All other products are my own.)

---

Enjoy reading The Pen Addict? Then consider becoming a member to receive additional weekly content, giveaways, and discounts in The Pen Addict shop. Plus, you support me and the site directly, for which I am very grateful.

Membership starts at just $5/month, with a discounted annual option available. To find out more about membership click here and join us!

One of the most enduring traditions in the U.S. is undoubtedly the state fair. The very first was held in Syracuse, New York, in 1841, and throughout the mid-19th century, states launched their own unique takes. Some of the largest and busiest, such as those in Minnesota, Iowa, and Wisconsin, have been running just about as long as the states have existed. And it’s no coincidence that some of the most well known and beloved events, which usually take place in the late summer or early autumn, represent the nation’s agricultural heartlands.

The exhibition State Fairs: Growing American Craft at the Smithsonian American Art Museum’s Renwick Gallery celebrates the unique crafts and customs of these annual festivals. From cows sculpted out of butter and 4-H contests to seed art and crazy foods, ingenuity is as much a part of the tradition as showcasing time-honored farming practices, rural culture, industry, and of course, arts and crafts.

The Minnesota State Fair is particularly well known for its seed art, which consists of tableaux made entirely of grains and seeds that are cultivated in the state. Artists often translate famous artworks using a variety of crops, and portraits of celebrities are another popular subject. When submitted for consideration, artists must supply a “legend card” that demonstrates which seeds they used.

The Renwick Gallery show highlights more than 240 objects and artworks, from elaborate contemporary jewelry pieces to historic quilts to a rhinestone rodeo outfit. A contemporary mixed-media resin necklace by artist and jewelry designer Morgan Hill celebrates treats and pastimes, while a geometric weaving by Agueda Martínez (1898-2000) highlights the extraordinary craft of Southwest tapestry blanket weaving.

Growing American Craft is the first large-scale survey of state fair creativity of its kind, representing 43 states and Tribal Nations. “Ribbon-winning artworks and engaging craft demonstrations illuminate the lives of the artists—their families, memories, honors, and struggles,” the museum says.

Part of the show’s aim is to highlight the pride, resilience, and inventiveness of rural communities, dashing stereotypes in the process and illuminating how fairgrounds are woven into the social fabric of the U.S. Visitors encounter a wide range of unique interpretations of state fair events, such as a pyramid of more than 700 glass jars of preserved fruits and vegetables by canner extraordinaire Rod Zeitler.

Crops themselves take center stage in many of the works, including an elaborate sweetgrass basket by South Carolinian artist Corey Alston, who draws on the Gullah Geechee basket traditions to weave remarkably imaginative designs. And Justin Favela has also reimagined the palatial Rubenstein Grand Salon into an immersive space titled “Capilla de Maiz (Maize Chapel).” The walls are fringed with gold and piñata corncobs, tapping into the role of maize in America throughout several millennia.

State Fairs: Growing American Craft continues through September 7 in Washington, D.C. The Smithsonian has also produced a beautiful catalogue to accompany the show, which you can find on Bookshop.

Do stories and artists like this matter to you? Become a Colossal Member today and support independent arts publishing for as little as $7 per month. The article ‘State Fairs: Growing American Craft’ Surveys Agriculture, Ingenuity, and Community appeared first on Colossal.

This is a guest post from Fred Hopengarten, K1VR
Past Director, New England Division.

The ARRL does a lot of good work. It fights for spectrum. It helps organize public service. It supports education and licensing. You should be a member. You should support the League.
And yet: membership is shrinking, the mood is frayed, and the organization could use a little oxygen. Maybe a lot of oxygen.

This year, four Division Directors may seek reelection: K0AIZ in the Midwest, N8SY in the Great Lakes, AC0W in Dakota, and K5UZ in Delta. They may be fine fellows. They may be good operators, decent DXers, repeater builders, public-service regulars, or all of the above. Good for them. But that is not the job. Directors are not hired to be pleasant at dinner. They are there to make decisions that strengthen the League and amateur radio.

On that score, I think these four have backed the wrong plays.

A League That Likes the Curtain

Start with secrecy. In the old days, any Board member could ask for a roll-call vote. That changed. Now a recorded vote requires a motion, a second, and five directors willing to say yes.

Maybe that sounds tidy. Maybe it sounds efficient. It is neither especially tidy nor especially efficient. A roll call takes less than a minute. What the new rule really does is make it harder to see who voted for what.

That matters, because some of the most important motions are the least glamorous ones: dues, governance, and League positions on public issues. If members want to know how their directors voted, they should not need a treasure map.

K0AIZ, N8SY, AC0W, and K5UZ all supported that change.

Campaigning, But Make It Tiny

Then there’s the election rule that says campaigning and fundraising must stay entirely inside the division holding the election. No outside support. No GoFundMe. No help from a friend, a club, or a brother in another division.

That is a strange way to encourage competition.

Think about the math. If a division has 10,000 members, even a modest mailing gets expensive fast. A challenger is supposed to build a campaign, spread the word, and overcome the built-in advantages of incumbency — but only with one hand tied behind the back.

It starts to feel less like election reform and more like incumbency insurance.

Residency, Or Else

There is also the residency rule: a candidate must live full-time in the division. Not just be based there. Not just keep a foothold there. Full-time.

That sounds neat until you imagine real life. People own cabins. People snowbird. People split time between places. People travel. People have jobs, family, and call signs that do not always fit into a clean little box.

But under this rule, a vacation place outside the division can become disqualifying. A fishing cabin can become a scandal. A winter house can become a problem. It is hard not to wonder whether this is really about eligibility or merely about protecting the people already inside the tent.

Don’t Mention the Other Guy

One of the more remarkable rules says a candidate may not make any comments, direct or indirect, about an opponent in campaign communications.

That is a campaign rule that somehow dislikes campaigns.

By that logic, challengers are supposed to run without contrast, without comparison, and without saying why the incumbent should be replaced. In other words: please compete, but gently, and never in a way that causes a voter to notice there is an incumbent at all.

That is not healthy democracy. That is upholstery.

No Real Megaphone

The communications restrictions are just as odd. Candidates may not use communications vehicles that carry campaigning outside their division, except for personal websites or webpages.
Meanwhile, incumbents have the League’s machinery, the visibility of their office, and the ability to email members in ways challengers cannot match. Think of it as a League halo. That is not a level field. It is a tilted one.

And the burden is not theoretical. Email lists are not handed to challengers. Postal mail is expensive. Club networks are restricted. Large national clubs can become unusable for campaign outreach. If you are trying to run against an incumbent, the system seems designed to make sure your message arrives late, thin, or not at all.

The Bigger Problem

The deeper issue is not any one rule. It is the accumulation of rules that make challenge harder, scrutiny weaker, and change less likely.

ARRL membership has been declining for years. The League needs fresh thinking, not procedural armor. It needs directors who ask whether a rule grows the organization, broadens participation, and encourages competition. It does not need a governing style that treats every awkward question like a threat.

Ask these directors what they have done to reverse the decline. Ask whether they voted for the rules that narrow debate, limit fundraising, and shield incumbents. Ask whether that is really the best way to rebuild trust.

Because if the answer is “we made it harder for challengers to compete,” that is not a reform. That is a warning label.

What To Do

The good news is that small elections can still matter a lot. In the Great Lakes Division in 2023, N8SY defeated W8CI by only 152 votes out of 4,198 cast. In Midwest, Dakota, and Delta, there were no challengers at all.

That should not happen.

We need candidates in the Midwest, Great Lakes, Dakota, and Delta Divisions. We need energetic, eligible, forward-looking hams who want to serve, not just occupy a seat. Public service people. POTA people. SOTA people. Digital folks. Builders. Volunteers. People who care about growth.

So here is the ask: find one good candidate in your division. Encourage that candidate. Provide support. Help them run. Make sure you are an ARRL member. And when the ballot comes, use it.

Because the League does not need fewer challenges to prevailing thought. It needs more.