Jeff Zakrzewski 🇨🇦

346 posts

Jeff Zakrzewski 🇨🇦 banner
Jeff Zakrzewski 🇨🇦

Jeff Zakrzewski 🇨🇦

@jeffz

Serial entrepreneur. Lover of all things Scalable and Data.

Toronto Katılım Temmuz 2007
693 Takip Edilen421 Takipçiler
Jeff Zakrzewski 🇨🇦
@ahmednadar That's what I thought as well, but when they reach out to them, they will reply and tell you it needs to go to 311. And then 311 will actually action these reports and assign investigators - I'm speaking from experience having created the sign report site!
English
1
0
0
6
Ahmed Nadar
Ahmed Nadar@ahmednadar·
@jeffz Illegal signs fall under MLS (Municipal Licensing & Standards) in Toronto, they handle bylaw enforcement including unauthorized signage.
English
1
0
0
33
Ahmed Nadar
Ahmed Nadar@ahmednadar·
Hey #Toronto، solveto.ca just got a major upgrade. Report any city issue in 30 seconds. Snap a photo, AI writes the report, it goes to 311 + your councillor simultaneously. Live map. 24 categories. Ward filters. All free. No app download. No account needed. 🧵
English
55
72
624
80.2K
Ahmed Nadar
Ahmed Nadar@ahmednadar·
@jeffz It is there. Give it a try it and adda report an issue, it works like a native app.
English
1
0
1
24
blah
blah@waleedkasim·
@grok @AtnsXBT No it is not please listen to the muaic in that 13 seconds clip and then let me know again.
English
2
0
0
27
AtnsMDX
AtnsMDX@AtnsXBT·
things nobody told guys about growing up: - the friends you thought were forever become people you check on twice a year - you start having your dad’s exact thoughts - you stop being invited slowly - the music you grew up with starts playing in supermarkets
English
89
410
10.7K
891.6K
Jeffrey Emanuel
Jeffrey Emanuel@doodlestein·
@jeffz Oh, I'm sure a lot of it was waiting for Rust builds and long test suites. Otherwise it would have hit usage limits. But that's the reality of coding in a language like Rust, and I don't see a good reason why it shouldn't count.
English
1
0
2
589
Jeffrey Emanuel
Jeffrey Emanuel@doodlestein·
I've seen a lot of skepticism recently from people about the claims around how long agents like Claude Code w/ Opus 4.6 can go autonomously without some kind of automated loop feeding them new instructions. People think it's BS because they haven't observed it. It's real. 17+hrs
Jeffrey Emanuel tweet media
English
49
15
258
28.8K
Jeff Zakrzewski 🇨🇦 retweetledi
Tarun Sachdeva
Tarun Sachdeva@tarunsachdeva·
Introducing Traces A new way to share and discover traces from coding agents, and a small step to make AI more multiplayer. Here's how it works:
English
51
52
431
51.3K
Jeff Zakrzewski 🇨🇦
@kurtbuhler I've had to tell people that I will pay for their 1st month subscription if they don't find Claude Code substantially better than whatever else they were using. So far, I haven't had to pay out anything. Try Claude Code.
English
0
0
2
108
Kurt Buhler
Kurt Buhler@kurtbuhler·
I've introduced ±2 dozen people to Claude Code in the last month or so. 100% of those people : - Agree it is better than GitHub Copilot or other agents - Find it impressive and easy to use - Are still using Claude Code on a regular basis Try Claude Code.
English
5
0
19
1.4K
Jeff Zakrzewski 🇨🇦
I’m surprised that article didn’t mention the removal of the 100% tariffs to vehicles coming from China. Which takes effect the end of March. Strategically it would make sense for Tesla to try and occupy as much of the 49K vehicle quota as possible, so I’m sure that’s what they are doing.
English
2
0
19
1.8K
Sawyer Merritt
Sawyer Merritt@SawyerMerritt·
NEWS: Tesla is delaying Model Y Performance deliveries for many people in Canada to late spring or summer (from March), with some now being told to expect their vehicles as late as July. driveteslacanada.ca/news/tesla-mod…
English
63
28
501
66.4K
Jeff Zakrzewski 🇨🇦
I just found the next generation Strawberry test for multimodal models. They can NOT understand word orientation in images - they do not seem to have spatial sense when it comes to words. I asked Opus 4.5 why it was mislabelling words as being at the wrong orientation and it said: ----------------- As for why I made the mistake - I think it comes down to a few factors: 1. Semantic expectation overriding visual analysis: I "know" that legends are typically presented horizontally at the bottom of documents, so I may have processed the content of the legend (recognizing what the words say) without properly attending to the orientation of the glyphs themselves. 2. Text recognition vs. orientation detection as separate processes: It seems like I can successfully read/OCR text at any orientation, but that reading process doesn't automatically preserve or report the orientation metadata. I recognized "LEGEND" and "CONTROL PANEL" without properly tracing whether the letters were stacked vertically or laid out horizontally. 3. Anchoring to the page frame: The legend is at the bottom of the rectangular image, which might have led me to associate it with "normal" bottom-of-page horizontal text. ----------------- The prompt I used (and will now use to test new models): Please extract a list of words in this image, and their orientation on the page. Assume most words are at 0/90/180/270 degrees, so snap to the closest angle based on that. Count the words at each angle, and list the angles in descending order of word counts.
English
0
0
1
28
Jeff Zakrzewski 🇨🇦
@svpino Even more importantly - you can pick your desired code quality level with all the LLM tools. Critical core system? Put all the extra checks and balances in. Experimental feature? Go fast and vibe code it. HUGE power & efficiency gains with this flexibility.
English
0
0
0
15
Santiago
Santiago@svpino·
I was talking to a guy who vibe-coded a saas, trying to explain to him what a huge mess the code is. I thought I had him. He looked at me completely confused. “So?” He said. That’s when it hit me: Vibe-coders don’t give a shit about code. They don’t understand it, they don’t touch it, they will never have to deal with it. They care about bringing their ideas to life and nothing else. They are 100% unburden by what many of us consider a showstopper. There’s an important lesson here.
English
1.2K
285
6.8K
825.3K
Brian Roemmele
Brian Roemmele@BrianRoemmele·
We won’t see it like this until 2042. The last full moon of 2025 reaches its peak on Thursday, Dec. 4. It is also a supermoon—meaning that it will appear slightly bigger and brighter than usual. But not quite like this alignment for a long time. Look up…
Brian Roemmele tweet media
English
54
36
448
20.1K
Jeff Zakrzewski 🇨🇦
@dieworkwear Oh let me tell you a story about: Maiden Name: 8wZVLab!a*MvbLWhq!^C!!YUAd First Pet: 3!WeRgHKdfRGw+-CyPk_khh!n Street: RvVaFR.Z*MbyH77Geedm7F.6y
English
0
0
0
163
derek guy
derek guy@dieworkwear·
tell me the lore behind your mother's maiden name, your first pet's name, and the street you grew up on
derek guy tweet media
English
469
1.1K
21.6K
381K
Jeff Zakrzewski 🇨🇦
@iamaheron_ It's a total game changer. I get 3-5 unsolicited calls a day, and this is the first time in years my phone has been quiet and usable. Extremely happy with this feature.
English
1
0
7
2.3K
ThePrimeagen
ThePrimeagen@ThePrimeagen·
I know I made a joke earlier but this is seriously concerning. I am not convinced it's ai, I think it's just accelerating this trend. The enshitification has been marching forward for years now and I hate it I know the rage is all about how you should do no typing, you should only use AI, and only luddites say anything different, but I urge you to practice programming and becoming an expert It is worth it
Wes Bos@wesbos

What the HECK is going on with tech? In the last week: Multiple cloud outages, x DMs totally broken, antigravity doesn't work, my watch is showing me 15 year old cal events, mac OS is a mess, email is spammed to hell and every nerd on here is talking like new AI is the second coming

English
243
365
6.1K
886.6K