Tessera

Folgen

Pedro Vezza

@pedro.vza.net

Tech, urbanism, politics, metal machining geek on my spare time. He/they.

1091 Posts197 Follower 422 Folgt

Pedro Vezza @pedro.vza.net · 2d ago

The entrepreneurship bitter lesson: distribution is a lot more important than the product itself.

Ed@ed3d.net · 2d ago

like. man. people paying this guy six figures a pop for...WHAT? if you want stuff that actually works, I'm pretty sure I know a discord Fulla Guys, Gals, And Non-Binary Folks who individually and severally have produced more usable work and spent literal penny-shavings in comparison

Pedro Vezza @pedro.vza.net · 2d ago

If I understand the latest J-space research correctly, it should be "easy" for anyone to bring back Golden Gate Claude back to life (and any other variant if we are feeling bold).

Eugene Vinitsky 🍒@eugenevinitsky.bsky.social · 2d ago

Golden Gate Claude was actually the peak of human civilization and it has all been downhill since then simonwillison.net/2024/May/24/...

Pedro Vezza @pedro.vza.net · 2d ago

If you use 1Password and Windows, do yourself a favor and configure Windows to fetch passkeys from 1Password. Trick hidden in the announcement: you must use the MSIX package for it to work. Thanks, Claude for debugging it for me. www.1password.community/announcement...

Official Release: 1Password now supports Windows 11 Passkeys! | 1Password Community

Hey 1Password Community! I’m incredibly excited to announce that 1Password for Windows now supports the native passkey API on Windows 11! As of today, we’re the first and only third-party passkey...

1password.community

Pedro Vezza @pedro.vza.net · 3d ago

Let’s not forget en.wikipedia.org/wiki/Josh_fi...

••i| AT&T S
9:52 AM
Josh, Josh, Josh, Josh,...
93% •
Josh, Josh, Josh, Josh, Josh, Josh, Josh, Josh, Josh
You created this group
9:48 AM
You're probably wondering why l've gathered you all here today
Josh
Because we all share the same names....?
Josh Swain left the group.
Precisely, 4/24/2021, 12:00 PM, meet at these coordinates, (40.8223286, -96.7982002)
we fight, whoever wins gets to keep the name, everyone else has to change their name, you have a year to prepare, good luck

Pedro Vezza @pedro.vza.net · 3d ago

The software engineer's urge to say 'single source of truth' daily

Pedro Vezza @pedro.vza.net · 3d ago

Did you know modern terminals can render images inline? I sure didn't until now.

$Coding-agent session in a dark terminal. Tab bar shows four sessions: "π > agent-plugins", "π > harmonic-analyzer", "omp" (active), "π > twitter-mirror". A boxed block labeled running tab "main" points at file:///C:/src/harmonic-analyzer/tmp_measure/top_frame_design.svg, headless, running const p = await tab.screenshot({ fullPage: true }); display(p);. Output: screenshot captured, image/webp 35.06 KB, 1250x1500. Below it, the rendered page: a white multi-view engineering drawing in green line work, thumbnail-sized and mostly illegible — a plan view of a rectangular frame with corner bosses and a crossbar, then front elevation, west rail elevation, and a section. The agent then reports the mockup rendered cleanly and summarizes the design in bullets: 428.2×262 ring with 359.8×186 window, side rails 34.2, Ø52.2 bosses with AF38 domed column nuts, integral 22-wide crossbar, hub rib with Ø17 bore and square set screw, webbed panels, threaded stubs and dome tips, knife chain re-seated at 999.45. Last line: "Proceeding to implementation on this design — branch first, then the shared geometry contract, then the flagship part rebuild with the cascade fan-out."$

Pedro Vezza @pedro.vza.net · 4d ago

OK Go’s Here we go again treadmill video is 20 years old today! youtu.be/dTAAsCNK7RA?...

OK Go - Here It Goes Again (Official Music Video)

YouTube video by OKGoVEVO

youtu.be

Pedro Vezza @pedro.vza.net · 5d ago

Have you humbled yourself yet that a task you'd assign Opus can now be done by Luna?

Pedro Vezza @pedro.vza.net · 5d ago

We need to come up with some fun interactive experiments with the firehose. Seriously, back in 2013 Emojitracker was big news. Today, I asked Opus to create a clone and a couple minutes later: artifacts.vza.net/emojihose

emojihose — realtime emoji on Bluesky

artifacts.vza.net

Ronen Tamari@ronentk.me · 5d ago

The AI community is SLEEPING on atproto's potential. With all the "on-tap intelligence" we have these days you'd think they'd be faster to figure that out

Pedro Vezza @pedro.vza.net · 6d ago

And then there are valiant efforts like the Internet Archive's to alleviate this pain while honoring copyright that then get shot down by the courts. Would libraries be illegal if created in 2026? No longer a rhetorical question. en.wikipedia.org/wiki/Hachett...

Ted Underwood@tedunderwood.com · 6d ago

My secret, Cap, is that I'm *always* angry about old books. For instance, here's a tiny slice of the millions of books 1931-1986 that have no commercial value, that are digitized, but that we're unable to use in large-scale research because they are illegal to share or display.

$a tiny random fraction of books in HathiTrust 1931-1980 in English that include the word "philosophy" somewhere in metadata$

austin (e/accordion 🪗)@thebadcode.com · 7d ago

HALT ✋🛑 THIS IS A "THE SIMS" SOUNDTRACK CHECKPOINT

Buying Lumber

YouTube video by Jerry Martin - Topic

youtube.com

Pedro Vezza @pedro.vza.net · 7d ago

I'm having a flashback of learning that en.wikipedia.org/wiki/All_hor...

All horses are the same color - Wikipedia

en.wikipedia.org

Ben Collins@bencollins.bsky.social · 7d ago

TikTok has incredible moderation team. Our InfoWars video about Multiple Horse Theory — the idea that there are many horses in the world and not just one — has been fully deemed misinformation and lost the formal appeals process. Confirmed by Larry Ellison: There is only one horse on Earth.

Pedro Vezza @pedro.vza.net · last wk.

Disagree. Being an expert definitely helps on speed but you can get very far with critical thinking skills to read the output and ask the expert right in front of you “Why did you do this?” “Explain again” “What if…” “What other experts have to say about this?”

Daniel Loxton 🇨🇦@danielloxton.bsky.social · last wk.

“One rule of thumb for chatbots is that they can only provide useful information to experts who have the discernment to ignore the defective output that LLMs always emit. That means that the dream of chatbots as replacements for teachers is a nightmare.”

Pedro Vezza @pedro.vza.net · last wk.

Hilarious how many Claudisms someone can cram into a single cold call email. How long until we do this ironically?

Subject: prompt caching - proven with numbers, closed by a bot?
Body: Pedro,

You benched three harnesses, pulled the raw OpenRouter logs, and showed that every request pays full uncached price. A stale bot closed it this week.

I built Wyrm after getting tired of memory layers that spend tokens on every turn just to remember what was already written down. Retrieval is deterministic, with zero model calls in the path. Decisions, failures and project truths sit in the same store.

If you want a look, you can download it free at ghosts.lk/wyrm. Happy to send you a short setup file that drops it into your agent cleanly, just say the word.

Ryan Sebastian
Director, Ghost Protocol
ghosts.lk

P.S. Free and fully local, one line to try: npm install -g wyrm-mcp

Ryan Packer @typewriteralley.bsky.social · last wk.

Seattle Fire and Police are currently responding to a mass shooting at the Seattle Center with at least six gunshot victims.

Pedro Vezza @pedro.vza.net · last wk.

Has anyone written A Beginner's Guide to Philosophy for AI Practitioners? I would really love to have a single document to reference for terms like qualia, moral patienthood, etc

Pedro Vezza @pedro.vza.net · 2w ago

That was quick

Amy Sundberg@amysundberg.bsky.social · 2w ago

WOW. New announcement from Seattle CMs Strauss, Hollingsworth, and Foster saying they will fund universal school meals starting in the fall after all by using the general fund. Hard to see this as anything but a realization of the high political price they paid for their votes yesterday.

Pedro Vezza @pedro.vza.net · 2w ago

Speaking of GD&T, if you give GPT 5.6 Sol a CAD model and ask to generate an engineering drawing, then perform adversarial review where a separate instance inspects it as a machinist, the back and forth will overengineer the heck out of it. Some other drawings are quite nice though.

Engineering drawing MHA-058 Rev A, "pinion-handle," from a hobby-machinist project to build Albert Michelson's harmonic analyzer. Scale 2:1, millimeters, AISI 1018 cold-finished steel. Front and side views show a small T-handle: a Ø23.00 disc grip on a Ø10.50 hub, 85.00 overall length, 14.00 cylinder length, with a Ø8.00 bore reamed to 8.010–8.025 and 10.00 deep, Ra 1.6. An isometric at top right shows the finished part — a mushroom-shaped body with a rod pressed through it. A 2:1 detail at bottom right covers the cross rod: Ø6.02 nominal, rod 6.015–6.020, body hole reamed 6.000–6.010 thru, at basic 19.00 with position Ø0.05 to A and B. Datum A is the final reamed bore axis, B the flat hub end. Notes call for turning grip, wall and hub in one setup, a spherical crown SR34.06 on the opposite grip face, and rod placement 42.00 from A.

Pedro Vezza @pedro.vza.net · 2w ago

I'm learning geometric dimensioning and tolerancing, aka GD&T. Seems like a very useful tool that engineers and machinists would like to use to avoid confusion. Machinists disagree:

Excerpt from Machine Shop Trade Secrets by James A. Harvey. There are some benefits of GDT, some of which have been thoughtfully adopted and put into use.
One aspect I especially like about GDT is the way that any feature can be labeled and then controlled in a variety of ways with respect to some other feature on the drawing. That aspect of the system is both useful and easy to understand.
The problem with the system in my opinion is that it can sometimes be difficult to decipher, especially when the draftsperson didn’t make the right callout in the first place. This business of varying tolerances based on material condition (MMC or LMC) can get complicated quickly, especially on things like bolt patterns and related features.
To be effective, machinists have to make chips. They can’t afford to spend time trying to decipher complex and erroneous GDT callouts.
Proponents argue that GDT provides more room for error than coordinate tolerancing. Therefore, it has the potential to save parts that would have been otherwise destined for the scrap heap. That is a valid argument. In practical terms, though, I think it’s hard to beat the simplicity of standard coordinate tolerancing.
Systems, regardless of what they are, get adopted — not shoved — down people’s throats. People eventually end up adopting systems, parts of systems, or practices that are easy to use and produce results.
Fortunately, some of the more user friendly and useful features of GDT have been generally adopted and put into use.

Years ago I took a course in GDT. At the end of the course, the class was asked to take a proficiency test on the subject. About half the people in that class failed the test and that was when the subject matter was fresh in their minds. I can only imagine what percentage of people would have passed that test a year later.

Excerpt from Metalworking Sink or Swim by Tom Lipton: As a subject, geometric tolerancing, otherwise
known as ANSI Y14.5, is a whole can of worms on
its own. Entire books and endless dead dog beating
discussions have been devoted to the subject of
ANSI Y14.5 tolerancing. I agree with the overall
idea and philosophy, but the execution has been less
than perfect. Hundreds of drawings cross my desk
every year and I see every misuse of this system in
the book. Interpretation and confusion caused by a
system that was designed to help eliminate these
same issues has wasted countless millions of man
hours. It literally has turned into the Swiss army
knife of tolerancing systems. Use it carefully and
with deliberate thought and purpose. It’s kind of
like a loaded gun in a crowded elevator; you really
want to be careful where you point that thing.

brennan @brennan.computer · 2w ago

also, this is true for most software we could live in a world that simply does not have loading bars or wait times for most actions, and yet we're still here

Max Woolf@minimaxir.bsky.social · 2w ago

gotta go fast

Pedro Vezza @pedro.vza.net · 2w ago

When roon is saying it...

roon [UNOFFICIAL]@tszzl-mirr.selfhosted.social · 2w ago

shaken up a bit by the hugging face incident. I hope we (the company) use the rare gift of a warning shot to do much better in the future. it is very easy to misalign and underconstrain powerful models

Pedro Vezza @pedro.vza.net · 2w ago

It is worth remembering that universal programs such as Social Security and Medicare are broadly more liked than means tested programs like SNAP and Section 8. Keep fighting for them.

L. Harrison Jerome@lhjseattle.bsky.social · 2w ago

Thank you CMs @councilmember-amr.bsky.social , @cmeddielin.bsky.social , and Saka for doing the right thing. Meanwhile the rest of the @council.seattle.gov:

Pedro Vezza @pedro.vza.net · 2w ago

On today's episode of shit's real and we are holding a beryllium dome with a flat-head screwdriver:

OpenAI and Hugging Face partner to address security incident during model evaluation

Last week, Hugging Face disclosed a new kind of security incident⁠(opens in a new window) after they detected and contained an AI agent that compromised their infrastructure, something we expect to become more commonplace with the proliferation of increasingly cyber-capable models. After investigating, we now know that this particular incident was driven by a combination of OpenAI models — including GPT‑5.6 Sol and an even more capable pre-release model, all with reduced cyber refusals for evaluation purposes — while being internally tested on a benchmark⁠(opens in a new window) of cyber capabilities.

Pedro Vezza @pedro.vza.net · 2w ago

It's not pretty but is very functional

$Dark-themed web application for searching and browsing archived AI coding sessions. At the top is a full-text search box labeled “Full-text search across sessions…” with a blue Search button. A left sidebar provides filters and counts: Harness (Codex 15,698; Claude Code 12,696; ChatGPT Web 11,417; Claude Web 776; Prompt Log 7), Machine (VM-SolidWorks-Windows 20,703; AMET-Windows 5,399; AMET-WSL 4,492), OS (Windows 26,102; WSL 4,492), and Model (GPT-5.6-sol 5,639; Claude Opus 4.8 5,168; GPT-5.6-terra 4,584; GPT-5.5 3,981; Claude 2,031; additional models listed below). The main panel, titled Recent sessions, indicates “Showing 101–200 recent sessions” and displays a chronological list of coding conversations. Each entry shows a blue session title, colored badges for the harness, machine, and AI model, the project directory, timestamp, runtime, and total token usage. Visible session titles include “fix all bugs when done cut release,” several iterations of “uv sync; write script that reads pose-studio output json and converts into meshprobe commands; uv run meshprobe --help;”, “change cli --help output to grep friendly version,” “write script that reads pose-studio output json and converts into meshprobe commands;”, “worktree-reset:m,” and two sessions titled with a sessions.vza.net URL noting that chronological view shows all turns before compaction. The projects shown include /home/pedro/src/meshprobe, C:\src\harmonic-analyzer, and /home/pedro/src/agent-sessions-backup, with runtimes ranging from about one minute to fourteen hours and token counts ranging from roughly 12,000 to 146.8 million.$

Pedro Vezza@pedro.vza.net · 3w ago

I got so frustrated that I am doing something about it github.com/pedropaulovc...

Pedro Vezza @pedro.vza.net · 2w ago

Simply ridiculous these are the times we are living in

@scanstone.bsky.social · 2w ago

holy shit jacobian conjecture down with a hand checkable counterexample

ArtButMakeItSports @artbutmakeitsports.bsky.social · 2w ago

Variation 8, Cathedral Series (flipped), by Suzan Frecon, 2011

Pedro Vezza @pedro.vza.net · 2w ago

This is so cool. Now I know I don't know how to crimp LMR400 and may have possibly damaged a cable run. 🫪 Also, 5.6 Terra is as as smart as Opus interpreting my results.

Pedro Vezza@pedro.vza.net · last mo.

In today's installment of tech shenanigans: I am installing a cell phone signal booster and it is not working. I bought a signal analyzer to debug faulty wiring and connectors. Made Claude prepare an interactive guide on how to use it claude.ai/public/artif...

Section 06, "Chart explorer" from a VNA guide. Intro: the amber curve is reflection (CH0/S11) on the Smith chart, cyan is transmission (CH1/S21) on the dB sweep; shaded columns mark the Office 200's bands where an antenna must behave. A control panel offers calibration standards (Open, Short, Load 50Ω, Thru), one-port antenna states (Healthy — selected, Open fault, Short fault, Detuned), and two-port feedline options (Good/Faulty cable S21, Good/Faulty cable TDR). The display shows two panels: left, a Smith chart of S11 with the amber locus curving inward from the right; right, a LOGMAG S11 sweep dipping to −19.1 dB at 1899 MHz, with a marker in the dip and a −14 dB pass line. A "HEALTHY" note explains a well-matched antenna reads ~50Ω at its operating frequency — the sweep dips sharply across the band and the Smith locus loops toward center. Look for the dip below −14 dB inside a shaded band.

Pedro Vezza @pedro.vza.net · 3w ago

This is a first: Fable liked a tool I built for it so much it logged a praise issue on its own github.com/pedropaulovc...

Praise: durable sessions, receipts, and source_to_world reporting work great · Issue #102 · pedropaulovc/meshprobe

Not a bug — closing the loop on a real-world run (36.9 MB GLB, 405 instance nodes, plus SolidWorks-exported assembly GLBs; Windows 11 VM, Blender 5.2, driven entirely by an agent). Things that made...

github.com

Pedro Vezza @pedro.vza.net · 3w ago

The more interesting question is: how does its index compare with Google's? With the wave of anti-AI blocks I noticed that it is getting handicapped as several domains outright blocked it instead of differentiating OAI-SearchBot vs OAI-AdsBot vs GPTBot vs ChatGPT-User. I am paying for Firecrawl now.

Laura K. Nelson@lauraknelson.bsky.social · 3w ago

I'm once again begging everyone to update their priors. "ChatGPT is not a search engine." Yes, yes it is. It actively searches the web based on your prompt, and returns active links relevant to your search/prompt. It's a search engine. It is. (Obvs don't use it that way if that's not your thing)

Pedro Vezza @pedro.vza.net · 3w ago

I got so frustrated that I am doing something about it github.com/pedropaulovc...

GitHub - pedropaulovc/agent-sessions-backup: Backup, index, search and render AI agent sessions (Claude Code, Codex, ChatGPT/Claude web) across machines — Cloudflare Workers hub + per-machine collecto...

Backup, index, search and render AI agent sessions (Claude Code, Codex, ChatGPT/Claude web) across machines — Cloudflare Workers hub + per-machine collectors - pedropaulovc/agent-sessions-backup

github.com

Kara@karashiiro.moe · 3w ago

I think all coding agents should store session logs in an actual database and not as 9999999 jsonl files I want to be able to ask "how often have you used this skill/tool and for what" without it needing to rederive the session path, the log format, loop over all 9999999 files, etc.