the right way to use model capabilities is not to ship 10x more features to prod
it's to spend more time understanding your users, trying experiments, building prototypes, learning about things you don't understand so that you can ship things that actually work
My small programming tricks:
- ci and cicd commands that automate running CI and CI/CD
- make ci - same, but local in Docker. This is my pre-commit hook.
Inspired by https://t.co/AvLv7gI0Dl
Favorite use of Claude this week - https://t.co/CbttwBAzzg with access to CI/CD and asking it to analyze where time goes and optimize. 50% speed up on one pipeline and 35% on another. Took 2 minutes total to write a prompt and review a PR, and the whole team will benefit.
Threads blocked my account for absolutely no reason and there is no way to contact customer support. What an app, magnificent! But one less social media to doom scroll with :)
Though now I'm wondering if Google or Apple can randomly disable my accounts...
It has been a great honor to serve the Ukrainian people as the Minister of Defense.
Here is what our team managed to achieve:
1. Disabled Starlink access for Russian forces.
2. Took over a Ministry of Defense with zero budget, took a risk, reallocated funds from payroll from the end of the year, and effectively invested them in mid-strike capabilities, fiber-optic FPVs, low-cost reconnaissance, ground robotic platforms, interceptor drones, and deep-strike drones.
3. Launched "Logistical Lockdown", this cut off enemy logistics and initiated the isolation of Crimea.
4. Continued the funding program for the Drone Line.
5. Launched a support program for modern drone-assault units that rely primarily on advanced technologies in combat.
6. Introduced a 70% advance payment policy for procurements made via e-Points on the Brave1 Market portal.
7. Fundamentally overhauled the procurement system.
8. Procured thousands of pickup trucks, buggies, and ATVs for the military for the first time—and did so through open tenders.
9. The drone interception rate rose from 83% to 91%, and the cruise missile interception rate soared from 47% to 87%.
10. Contracted Patriot PAC-2 GEM-T missiles for the first time, and submitted an application through an EU loan to purchase PAC-3 missiles.
11. Launched a baseline drone supply system for brigades and corps.
12. Launched a massive grant program for manufacturers of explosives and missiles.
13. Initiated an unpopular but vital transformation of the military.
14. Conducted three UDCG meetings, where we successfully broke through the Russian information trap claiming our defeat, restoring partners' faith in Ukraine. This secured $40 billion in support announced for this year (excluding the EU loan).
15. Launched the mechanism to utilize the EU loan for our military priorities.
16. Found a way to scale cheap missiles against jet-powered Shaheds and signed a record-breaking contract.
17. Our domestic ballistics. We radically revised the technical specifications, maximized accuracy, and reduced the cost by 30%.
18. Signed a contract to procure Gripen fighter jets.
19. Collaborated with the military to plan and execute Operation Auchan, which halted the enemy's mechanized offensive for six months.
20. Opened up exports under the Drone Deal program to attract investment and scale up domestic defense-industrial complex (OPK) production.
21. Launched the Trophy Lab, providing partners with the opportunity to study captured Russian military technologies.
22. Launched the Defense AI Center A1 to accelerate the integration of artificial intelligence on the battlefield.
Thank you to my entire team for their relentless 24/7 service. A special thank you to my family for their patience.
Thank you all for your support!
I will continue to work toward the mission I originally brought to the Ministry of Defense: defeating the enemy through asymmetry, the speed of innovation, and the power of organization.
To be continued.
A short story about deferring tech choices to thought leaders:
Early days at Disqus (~2010-2012), we made several frontend choices based largely on what thought leaders were promoting at the time.
One example: there was a big movement toward "micro-frameworks." Instead of larger, well-tested libraries like jQuery, you'd stitch together tiny interoperable micro libraries (Ender.js was one). Disqus was an embeddable JavaScript app, so file size mattered. It fit our use case, so we went with it.
Then it went live, and we were serving millions of users. The reality of those choices became clear. Micro libraries meant that instead of one good semi-bloated library, you ran 6-7 smaller, less-tested, crappier ones. We burned a ton of cycles fixing bugs and covering corner cases when we could've been shipping product.
We made a few choices like this.
At conferences, I'd track down those same thought leaders and ask for advice. "I'm hitting problem X, Y, Z. How did you solve this?"
That's when I learned my lesson: they rarely had answers, because they'd never reached our scale. Their energy went into promoting new stuff, not running it.
You should know this has never stopped. It's happening right now with AI. It'll happen again with whatever comes next.
Do your own homework. Test a lot. Don't just go with what somebody tells you.
I want to thank everyone for all the support we’re receiving. There really is an incredible amount of it.
For me, the sacrifice of the people depicted on the helmet means more than any medal ever could - because they gave the most precious thing they had.
And simple respect toward them is exactly what I want to give.
This is a huge release from the DynamoDB team -- no more synthetic keys (O#${orderId}#${itemId}#${createdAt}) to jam a bunch of values together for indexing purposes.
Much, much easier to evolve your table + add new access patterns without needing to backfill a bunch of data.
https://t.co/6qIe75CSkS
«Приїжджає два премʼєр міністра на вокзал. Треба буде спочатку їх пропустить, то трошки почекать, поки вони пройдуть, а потім обичні люди»
Пряма мова з поїзду ;)
@dizzzmas so I think automating that part of my life would be really cool.
actually prepared grocery bags would also suit. I remember there was ovo service that just brought me healthy seasonal food for the week and I didn't care about selecting each item
app idea: you make a photo of your fridge and the app generate a grocery list with stuff you need to buy.
bonus: it also goes to your favorite grocery delivery place and drafts an order
what do you think?
@dizzzmas right, but I feel like approximate solution would work just fine.
each time I have to fill my grocery delivery list in the app, I'm bored and would rather do something else. I really don't care about exact groceries, just give me nutritious food that is enough for couple days