I spent years cooking for people who said, “Anything is fine.” It never was. So I’m building Mate’s Table—turning your taste into an agent, because a 4.7-star rating still has no idea what you want for dinner. matestable.com
1
8
829
Love this. Research is no longer the moat. Fortune 1000 pulled and scored in 3.6 seconds flat, off 100+ data sources stacked into one call. Now it's a line item with little charges for a week of desk research.
Introducing Jev + Firecrawl Alexandria. We pulled data on every Fortune 1000 company, then Jev analyzed it in 3.6 seconds. With 100+ data sources, there's so much more to work with. Alexandria is live now for anyone using Jev and beyond!
2
27
Like to see more real usecase of Jev in production. Intelligence without memory is just a very expensive intern. Beacon (open source) use Jev to score all agent sessions, keep what's reusable, drop the rest, and continuously build a shared history across your agent harnesses.
Another insane Jev use case! Jev makes it incredibly cheap to evaluate and classify agent runs at scale. And finally, someone open-sourced a self-improving memory layer that can put that capability to work across agent harnesses. It turns your agent sessions into a compounding knowledge layer, where every successful run can make future agents smarter across: - Codex - Claude Code - Cursor - OpenCode and 20+ more Beacon by @asymptotelabs continuously builds a shared history across your agent harnesses and uses Jev to identify the runs worth learning from. It then turns the best workflows, corrections, and debugging patterns into reusable skills. GitHub repo: github.com/Asymptote-Labs/ag…. (don’t forget to star it ⭐) Most agent runs are messy. They contain exploration, failed commands, dead ends, and one-off fixes that should never become permanent memory. So Beacon preserves the full session history, while Jev helps decide what should be promoted, reviewed, or discarded. The recording below shows this in action. Beacon found 579 sessions across 5 coding-agent harnesses and normalized them into one consistent history. From there, Jev surfaces the lessons worth keeping and makes them available across your agent stack. - A pattern learned in Cursor can carry into OpenCode. - A lesson from Claude Code can improve the next Codex run. Every successful run adds to the shared knowledge layer, making future agents smarter. If you want to dive deeper into Jev, I also wrote a breakdown of how it works. The article is quoted below.
1
32
McDonald's spent 8.5 billion dollars to rediscover what a plastic tube already knew.
McDonald's is spending 8.5 billion dollars to find out why families stopped coming. I can save them the 8.5 billion. They took out the playground. The plan, announced Wednesday at investor day. Better food, faster service, new equipment, AI tools, hospitality training, and a bigger push into chicken and beverages. They call it Next. The goal, they say, is making customers feel more comfortable. Wall Street loved it so much the stock fell 5 percent to its lowest since 2022. Over five years McDonald's shares are down 3 percent. Burger King's parent is up 14. The owner of Taco Bell and KFC is up 13. This is not a hamburger problem. This is a McDonald's problem. So let us talk about who used to eat there. Parents. That was the whole business. Parents at 5:40 on a Tuesday with three kids in the back and nothing thawed at home. You did not go for the burger. You went because your kids could run screaming through a plastic tube for 40 minutes while you sat with a coffee and experienced silence. That was the product. The food was the cover charge. The first PlayPlace opened in 1971 in Birmingham, Alabama, and for fifty years it did one job. It made small children demand to go to McDonald's. Then 2020 arrived, every play area in the country closed, and a great many never came back. The equipment came out during a remodeling program named Experience of the Future. The Experience of the Future turned out to be a touchscreen. Their own chief executive said he did not know if ball pits were in their future. Sir, you were not selling ball pits. You were selling forty minutes. So now there is no counter, no playground, and no reason for a seven year old to want to go. A screen, a window, and a bag. And for that we pay prices that require a moment of reflection in the parking lot. For thin, overcooked, oddly uniform pucks that taste like a memory of beef. I did not love them because they were good. I loved them because I was eight and there was a slide. And have you seen the new buildings. The old ones were unmistakable. Red roof, yellow arches, legible from a highway at 70 to a child who could not yet read. The new ones are a flat gray box with a small logo and a drive through wrapped around it. A regional claims office with a fryer. They removed the color, the characters, the counter, and the playground, and then hired consultants to find out why it no longer feels like anywhere. Which is why, when I have to eat fast food, I go to Chick-fil-A. Not for the politics. For the competence. They have led the customer satisfaction index for fast food eleven years running, and average 8.5 million dollars per location against McDonald's 4 million. More than double, while closed on Sundays. Fifty two days a year with the lights off, and still double. A seventeen year old hands you a bag and says my pleasure and means it about 70 percent, which is 70 percent more than a touchscreen has managed. McDonald's is spending 8.5 billion on hospitality training. Chick-fil-A earns 8.5 million per restaurant by just having some. So here is my consulting proposal, free of charge, on behalf of mothers everywhere. Put the playgrounds back. Not a nostalgia campaign. Not a retro cup. The structure. Tubes, slides, a door the kids cannot open from the inside, and a bench where a tired woman can sit. And understand what you are buying, because it is not lunch. A six year old does not know your food is bad. He has no basis for comparison and no functioning palate. He thinks the burger is incredible because he is standing in a castle made of tubes. That is the window. That is the only window you get. You catch them before the taste buds arrive, and thirty years later they drive past your sign, feel something they cannot explain, and pull in anyway. That is not a menu strategy. That is imprinting, and you shut it down in 2020 to save on cleaning. Eight and a half billion dollars for AI tools and hospitality training. Or a slide. 🦋
34
Feels like the next big platform tax gets collected at that door, not at checkout.
This is a big enough problem that there's not just one startup idea here, but several. For example, there will probably be one variant meant to filter out agents, but another meant to welcome them in an orderly way. This also seems an idea you could build more on.
41
Self-hosted, open source, and it runs on a free Cloudflare account. Five platforms out the gate, X, Threads, LinkedIn, Mastodon, Bluesky, with more coming. Every scheduler that charges monthly for this exact feature set just lost its reason to exist.
🎉 Introducing CogSend.com! An open-source social media scheduler tool that you can self-host on your free Cloudflare account. CogSend currently supports X, Threads, LinkedIn, Mastodon, and Bluesky with more coming soon. More details 👇
2
2
2
180
Another real usecase of Jev in production. Treg (OpenRouter for agent tools, open source, self-hostable) on Jev, run $0.0089 per lead instead of a $600/month subscription, subscription-vs-pay-per-use, open source, zero markup.
We rebuilt Clay for agents Powered by Jev + @treg_ai No more $600 subscriptions, just $0.0089/lead Try it at Treg.to/people-search - 85% cheaper than Clay - #1 on people search bench accuracy - Plugin to any agent Fully open source, 0% markup Git Repo below 👇
80
A library of 12 million shorts can tell you which hooks worked. It can't tell you why they'll work for your business next week.
jev has made the impossible POSSIBLE. so this morning my real estate buddy asked me how to break out on instagram. I could: - ask her to scroll. - find her favorite creators - map out each hook - figure out what makes them work. - rework it in her voice - rinse & repeat every single week. Or i could have jev go through a library of 12 million shorts and do it for her.
52
Google will patch the NotebookLM SEO trick exactly the way it patched the Gemini share-link trick. Every open door in a platform is a closing door.
black hat SEOs are leveraging public NotebookLMs being indexable by Google as a new parasite SEO tactic! in case you already don't know! google is indexing public Notebooks created in NotebookLM and a lot of black hat SEOs are leveraging this for parasite SEO cuz those public Notebooks are hosted at notebook .google .com and cuz of google .com authority these URLs are performing well too similar situation happened with Gemini chats too sometime back and Gemini team made some changes such that the public shareable links are not indexable by Google rather it's just like you can take it and share it with your friend or someone else and they can see the conversation that you had with Gemini and I feel like same is going to happen with these notebooks too where NotebookLM team is going to make changes in robots.txt or maybe no index all these public URLs by implementing a noindex meta tag on all these pages and it's just going to be that if you create a public shareable notebook, then you can just share it with others and it's not going to get indexed by any search engine
1
1
235
No body tested your vibing code live. Choose up to 999 toppings is the kind of copy that only ships when nobody actually orders a burrito before merging the branch.
It’s embarrassing to see engineers+designers at UberEats no longer looking at what they ship. Uber introduced a new UI to select toppings… except when there’s no limit on how many toppings to select, the placeholder is “choose up to 999”. Showing everywhere. QA nonexistent
1
34
In Australia, hospitality award rates provide waitstaff around AUD $26.44–$33.05 per hour. If you tip here it a reward for good service. While US model allows menu prices to stay lower and enabling many servers to earn competitive total pay through variable tips. That's the entrenched differences between the two countries.
My husband and I went out to dinner with four friends last weekend. Six people. The bill was around $150. The restaurant had already added a 20% service charge. I noticed it immediately. My husband didn't. We were talking and getting ready to leave when the waitress came back and stood beside our table. She looked at my husband. Then at the receipt. Then back at him. He looked confused. “Is something wrong?...
112
Two agents wasting tokens over a Verizon bill on human's behalf is either the funniest outcome of this whole AI wave or the most depressing.
Letting my AI agent negotiate against Verizon’s ai agent to get a discount on my internet bill
1
81
Spread. Two hundred thousand dashers underpaid for two years and nobody caught it until the city did the audit, that's not a glitch. What the internal dashboard was actually showing them?
Since 2023, more than 200,000 Dashers in New York City were underpaid or paid late under the City’s delivery worker pay rules. These errors were unintentional, but that doesn’t make them acceptable. Dashers should be paid in full, on time, every time. We take full responsibility, and we’re deeply sorry. We’ve reached a $131.5 million settlement with the NYC Department of Consumer and Worker Protection. Every impacted Dasher will receive the amount they are owed, and they’ll be contacted directly with payment details. We understand the impact of even one missed or late payment. We’ve addressed the issues that caused these errors and made changes to prevent them from happening again. Read more here: about.doordash.com/en-us/new….
Community note
theguardian.com/us-news/2026/s… The City of New York asserts this was not a mistake. “For years, DoorDash failed to count every hour worked by delivery workers,” Mayor Mamdani said. “This was not a rounding error or an accidental mistake.
23
Cutting the top seller instead of raising its price is the harder call, most places do it backwards and lose regulars either way.
The owner of a top BBQ joint in Texas decided to cut brisket, the top-selling item, from the menu rather than institute a $10 markup for customers. The decision was fueled not just by the skyrocketing price of brisket but by inflation. wapo.st/4h8FCdT
11
Same-store sales up 0.8 percent while beef and labor costs climb, that's the real squeeze today.
MCDONALD’S CEO: HIGH INFLATION IS HERE TO STAY McDonald’s CEO Chris Kempczinski says weak customer traffic and persistent inflation are becoming the new normal for the restaurant industry. U.S. same-store sales rose just 0.8% last quarter, while restaurant traffic declined. With beef, labor and construction costs elevated, McDonald’s plans to compete aggressively for market share, while remaining cautious about further menu price increases.
18
Close to the technology or close to the customer, and almost nobody is good at both at once, that's basically the whole org chart problem right now. Most AI startups are quietly failing because they hired for the wrong extreme.
A CS undergrad asked me where he could have most effect in the AI age. I said probably at either extreme: either close to the technology, actually making LLMs, or close to the customer, using AI to give them exactly what they want. Or maybe both if you can stretch that far.
11
Phone screen itself became the friction Meta wants gone. Shipping right in xmas season is either brilliant timing or a hardware launch nobody actually asked for.
Replying to @finkd
And we've been building a joyful little device for talking to your Muse that fits on your keychain. Muse Charm -- shipping in December.
24
Half these apps assume everyone already knows what a promo code is, and Nani just proved the whole UX team wrong in under five minutes.
This is so wholesome A granddaughter teaching her Nani how to use a quick-delivery app 😍
1
94
More is coming for real usecases of Jev in production. Feed vicaura a URL and get back features, ICP, pricing strategy, messaging, straight into a markdown file. That's a competitor teardown that used to take an afternoon, done before your coffee's cold. Research just got a lot less precious.
Replying to @mmmikhaeel
You can try it here vicaura.com
54
Read ten explainers and watched five hours of video just to understand a model that answers yes, no, or pick one. That's not a hard model. That's bad naming.
I read 10+ articles and watched 5 hours of content to understand what Jev actually is. This is my attempt to put it all in one place, from how it works and what you can build + list of useful resources needed to go deeper.
Article

Jev 101: what it is, and what you can build with it

I have read 10+ articles, watched 5 hours of content to understand what actually Jev is and then chose to write this. It is not a chatbot, It’s a whole new kind of AI, more like a classifier which

1
22
Every new restaurant fit-out now reads like the same three Pinterest boards, exposed bulbs, pale oak, one plant nobody waters.
Minimalist aesthetic needs to die Everywhere practically looks the same 😭 Frame 1, home. Frame 2, content studio Frame 3, restaurant
2
31