Craig Tataryn (@craiger) header
Craig Tataryn (@craiger) profile picture

Principal Developer | Grind Software Inc. | Typescript Scala GraphQL

Winnipeg, MB, Canada
Joined March 2007
Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)Photo by Craig Tataryn (@craiger)
A BLEEP at the right moment can turn anything dirty! Whoever made this is a mad genius.
126
551
104
3,348
229,693
Craig Tataryn retweeted
Lmao
63
690
68
6,076
430,006
Craig Tataryn retweeted
"Do you use Jira?"
194
548
102
5,880
519,327
TL;DR: Anthropic copied the code that took human engineers over 37 years to build and it cost $20k + the cost of a researcher.
Anthropic had 16 AI agents build a C compiler from scratch. 100k lines, compiles the Linux kernel, $20k, 2 weeks. To put that in perspective GCC took thousands of engineers over 37 years to build. (Granted from 1987 - however) One researcher and 16 AI agents just built a compiler that passes 99% of GCC's own torture test suite, compiles FFmpeg, Redis, PostgreSQL, QEMU and runs Doom. They say they "(mostly) walked away." But that "mostly" is doing heavy lifting. No human wrote code but the researcher constantly redesigned tests, built CI pipelines when agents broke each other's work, and created workarounds when all 16 agents got stuck on the same bug. The human role didn't disappear. It shifted from writing code to engineering the environment that lets AI write code. I don’t know how you could make the point AI is hitting a wall.
3
99
Craig Tataryn retweeted
Today's free video on Matthew Explains looks at the "goldfish" loss function, used in the recent Apertus language models. It's supposed to prevent copyright infringement by making the model unable to reproduce word-for-word quotations. Does that work? matthewexplains.com/11182442…
1
2
142
I really hope the opportunity is missed for years to come. First of all it resides within a park. Cottages there are on leased land and are very modest, using a shared well to fill cisterns as needed for water. Sewage is hauled away by trucks because there's no infrastructure.
The beautiful white sand beaches on Lake Winnipeg feel like a missed opportunity. Like ocean beaches, with nothing similar for thousands of kilometers, they might be a destination from across the centre of the continent if developed with quality resorts and tourist amenities.
79
Craig Tataryn retweeted
My latest video just went live, and as well as being free, it covers some points I know are of interest to people here regarding the distinctions among training, fine-tuning, and prompting.  I'm hoping it gets a lot of views. matthewexplains.com/11179488…
1
2
190
I really wish more founders understood this. Also, the ability for the team to communicate effectively and understand requirements is key.
Replying to @sanjeevn72
absolute advantage of a cracked team in the age of ai! just not 6-9 hours apart, which leads to 24-hour long iteration cycles. ai made iterations even faster, making timezone difference even worse
53
Someone is feverishly trying to get an LLM to churn out some codecs 🤣🤣🤣
Replying to @FFmpeg
You clearly have no idea how much of my history was in ffmpeg. I built a ton of early twitch infra on top of yall. I’m not talking out of my ass. I’m complaining that a valuable piece of software can’t even play a fucking video recorded on an iPhone because “v4 soon”
113
Highly recommend, I’ve watched the GAN video and it was enlightening. Matthew does a wonderful job explaining indeed.
I've launched my latest project, a video service called Matthew Explains focused on machine learning and computer science. (link in thread for algorithm reasons)
1
76
Return to Castle Wolfenstein (multiplayer)
56
Craig Tataryn retweeted
What game?
47,829
2,343
12,147
91,501
21,194,969
Craig Tataryn retweeted
He is the creator of linux, the open source operating system which powers 96% of all servers on the planet. To facilitate the development of linux, he created a side project called git. Today, git is the most used version control software in the world. What an absolute legend🫡
149
164
27
4,642
130,344
Study by my genius friend to CURE DEAFNESS cancelled because, I shit you not: "Research programs based primarily on artificial and non-scientific categories, including amorphous equity objectives, are antithetical to the scientific inquiry". Decel anti-science brain-dead take.
My lab’s 5-year NIH R01 grant, awarded to study gene therapy for hearing loss, was abruptly terminated. I want to share how this action has been incredibly harmful and disruptive, not just to my lab, but to the scientific process itself. 1/15
48
203
32
3,612
758,817
Craig Tataryn retweeted
sorry not sorry but you gotta know this cursed SQLite fact too
79
280
51
3,923
264,504
Even before AI I’ve had PRs rejected because I added comments. The review stated “the code should explain itself”. Or “comments are a future lie”
you're not allowed to write comments in your code anymore, because if you do everyone will just think it's ai generated.
76
Craig Tataryn retweeted
More than $30,000 has been raised through $55 donations for the True North Youth Foundation following the passing of Mark Scheifele's father. Bigger than hockey 💙
30
320
24
4,772
132,525
Congrats @DallasStars, you put up tremendous pressure in our zone and it ultimately paid dividends. See you next year #NHLPlayoffs #GoJetsGo
81
Craig Tataryn retweeted
The Blues fans’ reaction videos to the Jets tying/winning goal are tough to watch 🥲 (via TT/toastedravs, TT/crystalandpenny)
279
386
136
5,626
1,123,853
They really missed an opportunity by not calling it CRIT
I’m an Incident Responder on the AWS Customer Incident Response Team (CIRT). And I get asked a lot of questions, like: “Where do I even start with incident response in the cloud?” Here’s a beginner-friendly thread on AWS IR tips — with a few lessons I learned 🧵👇
50