Phil (@parzerp) header
Phil (@parzerp) profile picture

co-founder at https://teapawt.org/t.co/V1Sa1tPVLr, implementing AI for DACH companies at https://teapawt.org/t.co/SyzwTAodYn

Vienna, Austria
Joined March 2014
Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)Photo by Phil (@parzerp)
I built my girlfriend a personal site for her birthday people love my own personal site so I decided to go with a similar style for hers, kept the storytelling + fun little animations
86
83
14
2,820
242,793
yapatron in hong kong
3
1
15
470
gm from hong kong
3
12
324
BREAKING: Famous Austrian restaurant critic @parzerp was recently spotted in Hong Kong 馃毃
1
8
295
just told my mom about my flight getting cancelled she drops a random sticker of my grandpa as a reply
3
8
236
china ecom ops are crazy
1
6
185
they should add an s tier to nutri score for the all time goats potato, raspberry, topfen, lean ground beef and egg
6
209
not much beats writing on a long train ride with no wifi
1
8
391
crack cocaine
1
9
282
Spent the weekend with two german CS academioids (phd) dire learnings: Claude is mainstream. For them ai == claude. Every other sentences is "just use claude for xyz", "prompt my claude...". They act shocked about my gf not using claude to plan our china trip??? They are on $20 plans, try to trashtalk me (and anyone else) for being on multiple 10x/20x plans. I explain tokens as a business expense, they kinda get it. They are happy that they don't need software devs anymore and seem amused by the fact that they never really wrote any code or learnt how to read it. They don't really know about open weight models, assume they are unusable and not worth investing any time in. Their gold standard and the only thing they know about re open weight is Gemma. They are still copy pasting "scripts" into and from claude desktop+mobile. They don't know about computer use or putting AI in a loop (e.g. testing and optimizing with computer use, feeding logs, etc) They don't take it seriously when I tell them about it ("AI can't reason like a human") They are big on skill files, just finding out about subagents and "checking and cleaning up their claude md daily" Using AI in any step of the design process invalidates digital media for them (I showed them my past work, first they praised it, when I told them where I used AI images/videos/graphics they started criticizing) They have started working on hobby projects and "scripts" from their phone. In general they are not allowed to use AI for work but there's "working groups" at their jobs that are trying to "figure things out" (= how to use AI "safely"). Cloning software is amazing (referencing Adobe product clones, specifically photoshop clones). I asked them why not just use an open source alternative, they talk about people wanting to get a job where Adobe Suite is a requirement??? Data Centers bad. Overall they don't think they'll be out of a job soon because "AI is just a tool". Also "AI is a stupid term for what this tool is, cause it's not intelligent".
5
1
13
994
bro starts leaking ui when im ooo for one day it's so over
tired: slide to merge wired: hold to debrief
1
4
287
Two categories of products worth building today: - Stuff you actually care about and use every day To meet today's bar you need to care, if you don't use the thing you are never going to care enough. - Hyper-ambitious, moonshot ideas Reaching for the stars is always valid, so if you have a grand idea and something validating it, it's probably worth pursuing this, even if you're not in the position to use the thing you're building (yet). Building some pre-AI ZIRP-type SaaS for a "niche" you just learned about and have no connection to is dead. Same goes for consumer products targeting an audience of consumers that you are not a part of.
4
9
398
Phil retweeted
hot take: you don't need a software factory, and i don't love the term. it's just marketing all you need is good software engineering and investing your tokens in the right places. the more deterministic the better (and token efficient)
all the companies selling you a software factory produce a shockingly tiny amount of software
439
583
180
8,135
510,879
if your coding environment isn鈥檛 optimized for walking around all day you鈥檙e missing out
3
3
15
2,134
lil bro just didn't know it was expensive to take screenshots at every click and pipe them into context, nobody told him :(
1
4
303
another fun german boomercast episode in the bag
1
5
246
Phil retweeted
Good art is dictatorial in nature The absolute worst thing you can do to a great creative idea is to democratize it, and get too many opinions on it, too many people working on it The best art comes from one person having totalitarian control, relentlessly pursuing their vision
13
20
4
150
8,363
what an incredible time to be alive I will never get the doomers
1
5
377
Phil retweeted
the human behind the code is the moat. you are the moat.
there is no moat
98
74
22
1,260
60,295
hope Europe feels safe with these two jesters defending it
2
1
7
532