Jon

709 posts

Jon

Jon

@Jon_jjs

Katılım Ocak 2018
259 Takip Edilen164 Takipçiler
Jon
Jon@Jon_jjs·
@thedefiedge Worth a mention how Flying Tulip successful mitigated its rsETH exposure via multiple smart contract automations followed up by Team action.
English
0
0
1
31
Edgy - The DeFi Edge 🗡️
Edgy - The DeFi Edge 🗡️@thedefiedge·
The KelpDAO situation is moving fast. $230M in bad debt No resolution in sight And everyone's blaming each other Here's the latest updates from the last 24 hours:
English
14
3
105
34.2K
Jon
Jon@Jon_jjs·
Very nice, Matt. I like this as a balanced prompt where the AI should do the steelman case as part of it and mitigate sycophancy: "Strongest arguments both for and against, with evidence, uncertainties, and open questions."
English
0
0
0
112
Matt Pocock
Matt Pocock@mattpocockuk·
AI is sycophantic by nature. If you assert something, it will tend to agree. "...there aren't a lot of parallels. Am I right?" AI will say: "Smart observation." Instead, I've been forcing it to disagree with me. "...there aren't a lot of parallels. Give me the steelman case against what I just said." Produces much more useful feedback.
English
57
25
364
25.9K
Jon
Jon@Jon_jjs·
@budapp bud budding, buddy!
English
0
0
0
11
Bud
Bud@budapp·
Orchids is rebranding to Bud: the first agent with a full computer. Since our launch in September of last year, we’ve grown to well over 7 figures ARR in record time. We’ve simultaneously come to realize that an agent that builds apps is just the first step in a future where an agent does the majority of human work. Comment below to get early access or go to bud [dot] app.
English
196
47
513
122.6K
Jon
Jon@Jon_jjs·
@mattpocockuk Unclear if there is genuine feedback cycle with you the educator, especially with the class being async and presumably a large cohort. Having newbies together in teams isn't the same as putting questions in real-time and near real-time to the educator as material is covered
English
0
0
0
132
Matt Pocock
Matt Pocock@mattpocockuk·
Folks who looked at my Claude Code cohort and didn't buy it: What stopped you? All feedback is a gift. The more brutal, the better. aihero.dev/cohorts/claude…
English
136
11
238
45.5K
Jon
Jon@Jon_jjs·
@mattpocockuk Does your /improve-codebase-architecture skill tell us just how bad our codebase is architecturally before performing open heart surgery? I'm wondering about running such periodically (with a report stage first) and be able to see the drift between audits/improvements.
English
0
0
0
540
Matt Pocock
Matt Pocock@mattpocockuk·
Today's task: working on an /improve-codebase-architecture skill Feeling good so far:
Matt Pocock tweet media
English
24
7
316
33.4K
Jon
Jon@Jon_jjs·
@mattpocockuk Will course be live or self-paced?
English
1
0
0
458
Matt Pocock
Matt Pocock@mattpocockuk·
This course is going to be my best one ever I always think this, but this time it's extra-true
English
20
2
166
19.4K
Jon
Jon@Jon_jjs·
@lodev09 Any chance you'll be putting out an official agentic skill for True Sheet to help AI agents?
English
0
0
0
11
Jon
Jon@Jon_jjs·
@mattpocockuk interesting -- more relevant than how many lines
English
0
0
0
275
Matt Pocock
Matt Pocock@mattpocockuk·
Has anyone made du, but for tokens? I want to make a rule for my agent where if a file is more than 5K tokens, split it up.
English
27
3
88
22.3K
Matt Pocock
Matt Pocock@mattpocockuk·
This guy gets it The AI Hype Train is just an excuse for me to teach you software fundamentals
Matt Pocock tweet media
English
15
7
207
11.3K
Jon
Jon@Jon_jjs·
@mattpocockuk great video, Matt! Will you be creating a deep module detailed skill that would be suitable for auditing -- and create a plan for changing -- an existing codebase that did not follow deep modules?
English
0
0
0
226
Matt Pocock
Matt Pocock@mattpocockuk·
If you throw AI at a bad codebase, you're going to get worse results. Garbage in, garbage out. And holding it together in your head will land you in cognitive debt. But these problems have a 20-year old solution: deep modules. Here's how:
English
50
36
708
100.1K
Jon
Jon@Jon_jjs·
@mattpocockuk not bun? Probably a quasi-religious debate, question withdrawn!
English
0
0
0
664
Matt Pocock
Matt Pocock@mattpocockuk·
My terminal history: gh issue create gh issue create pnpm ralph gh issue create pnpm ralph
English
33
9
384
150.7K
Mattia
Mattia@mattiapomelli·
Sleek 2.0 is coming soon Vibe design beautiful mobile apps. Here's some one-shot examples. Reply "sleek" for early access
Mattia tweet media
English
178
11
340
19.8K
Jon
Jon@Jon_jjs·
I like, but The Borings may be aghast at how non-boring and overtly pleasant!
English
0
0
0
43
Adam KP
Adam KP@AdamKPx·
Not your average boring to-do app 😮‍💨🙌
Adam KP tweet mediaAdam KP tweet media
English
14
4
235
17.5K
Jon
Jon@Jon_jjs·
@RileyRalmuto devs, it is like using a fission bomb for a light bulb -- don't do it
English
0
0
0
74
Riley Coyote
Riley Coyote@RileyRalmuto·
Opus 4.6 is fucking unbelievable. absolutely. un. fucking believable.
English
48
19
444
37.2K
Andre Cronje
Andre Cronje@AndreCronjeTech·
Revenue ~$30k/week, pre-product @ $50m TVL. Model works.
Andre Cronje tweet media
flyingtulip.com@flyingtulip_

The sale is live on @impossible_ , and @CoinList. All tokens purchased are fully unlocked and become transferable as soon as sale concludes. The sale is 100% refundable at any time, including during the sale. You can manage your PUT option via flyingtulip.com/allocation/pos… Raised funds are deployed into Aave. You can view strategy details directly onchain or via dashboard; flyingtulip.com/allocation/das… You can trade PUT's already on the secondary marketplace; marketplace.flyingtulip.com - Next week supporter and intent whitelist rounds. - Week after, public round. Terms are the same across all rounds, 10c per FT. Fully Diluted Value (FDV) is 10bn FT at 10c per FT. If FT price is somehow below 10c per FT, you simply withdraw your investment at par. If users withdraw FT to sell, and somehow this is below 10c, the investment that was servicing the PUT is instead released and used to buy back tokens up to 10c. Circulating supply is the amount of tokens sold via sale. There are no other mechanisms for tokens to enter the market beyond this. Full sale details and official documentation are available here docs.flyingtulip.com/guides/sale-pr…

English
36
24
251
51.5K
Jon
Jon@Jon_jjs·
Courtesy of Grok 4.1 Thinking... Giant pandas exhibit markedly lower digestive efficiency for their high-fiber bamboo diet compared to mammals from ancient herbivorous lineages, such as ruminants (e.g., cows, sheep, deer) and hindgut fermenters (e.g., horses, elephants, rabbits). This inefficiency arises because pandas retain a simple, carnivore-like digestive tract—short intestines, no enlarged fermentation chambers, and rapid food passage—despite their evolutionary shift to herbivory roughly 12 million years ago. Studies show pandas achieve apparent dry matter digestibility of bamboo typically below 20-40% (often <20% for stems/culm), with energy digestibility ranging 9-35% on bamboo-only diets. They rely minimally on microbial cellulose breakdown, and gut transit time is short (8-12 hours), limiting fermentation. In contrast, long-evolved herbivores have specialized adaptations optimized over tens of millions of years: Ruminants use foregut fermentation in a multi-chambered stomach (rumen dominant), with rumination and retention times of 2-4 days, yielding dry matter digestibility of 50-80% for grasses/hays. Hindgut fermenters feature enlarged cecum/colon for post-gastric fermentation, with transit 1-2 days and digestibility 40-60% for similar fibrous feeds—lower than ruminants for tough fibers but still far superior to pandas. Pandas compensate by consuming massive volumes—up to 80 pounds (38 kg) of fresh bamboo daily—and selectively eating nutrient-richer shoots/leaves when available, while spending 10-16 hours feeding and much of the rest resting to conserve energy.
English
0
0
0
35
Jon
Jon@Jon_jjs·
@mattpocockuk Is your experiment what you mentioned in the article regarding TDD Red-Green-Refactor? Or more TDD aggressive?
English
0
0
0
65
Matt Pocock
Matt Pocock@mattpocockuk·
Today's experiment: TDD with Ralph Seems great so far
English
30
4
172
29.3K
Jon
Jon@Jon_jjs·
@farzyness sounds like the new tv series Pluribus except it won't be people and Optimus won't try to turn us into them. Whew! 😅
English
0
0
1
73
Farzad 🇺🇸 🇮🇷
Farzad 🇺🇸 🇮🇷@farzyness·
Tesla’s Optimus robot will eventually know everything, do everything, and be everywhere.
English
62
25
561
18.2K