Indie Forward Deployed Looping Factory Engineer & TDDevelooper ✦ sponsor․sport ▶︎ ✦ dojo․foo ⏸︎ ✦ epicat․com ⏸︎ ✦ github.com/tomsiwik

Terminal
I don't care if you have 3 followers or 200k, I will follow you because your feed seems interesting to me
121
7
600
55,270
Yo @X / @Support am I shadow banner or my account bugged? My impressions don't match at all nor does it reflect correctly what those views are. Regardless of my content I'm always capping at around 50k. Why is that?!
8
11
683
It definitely is broken - how did this reach 61k views on a repost and wasn't credited at all to my post - wtf is this?!
33
Tom Siwik retweeted
Footer design for new project
1
1
45
Replying to @X @Support
Also, whenever I go on my profile and highlights, posts vanish. Some of my recent replies also appear on my timeline on my "For You" (why would I care). I seriously think my account is bugged or something. Don't know what's the reason for this buggy-ness.
57
Perfect analogy to settle this AI-assisted reading/writing code or let AI fully vibecode. Cooking. This AI vibe cult will tell you "just order in, bro" Order in: - costs $$$ - no shopping time - only few good restaurants - good taste (might get better) - call to complain if that not what you've ordered - closed and you'll starve Cooking: - costs $ - you can use precooked ingredients - shopping freedom (setup your harness) - might take longer depending on recipe - everything fully tweakable - best taste - closed but you can whip something up Who earns more? Cooks or fastfood servers? Exactly
4
22
1,472
This apathy towards code in social media is so concerning to me that I'm starting to believe people have developed serious psychosis using AI. Do me a favor and unsub/ pause your premium model for a week. Use qwen/gemma or some other local model & try to code something with it (even better: without AI) Anything easy, really. If you're struggling, don't you think you've atrophied a skill? Or do you seriously believe your way of thinking to abandon something so essential as merely reading code is healthy? It's similar to what addicts would do: bend reasoning / reality until it's "okay" to be an addict. "No, I can always stop". Even worse... pulling more people into this psychosis cult and encouraging terrible practices. Do your drugs alone please and leave the people be that still enjoy coding.
27
21
258
10,942
Unproductive sleepy people. Just get this app:
another 26.7M views for the alarmy app this is 🔥
1
6
1,377
The more believe this the better for me. I've run an experiment disproving this. Results tomorrow (drafted) Sneak-peak: would have costed me $ 2k w/o rw the code and $ 700 w/ reviewing + w/o writing the code Lost: 10 days (smart model) Won: 5 days (dumb model)
Tired: Not writing the code. Wired: Not reading the code.
2
14
976
∙ New Logo ∙ Shiki ∙ Twoslash ∙ Modern Monaco ∙ Slidev ∙ Vercel/eve ∙ Vercel/ai-sdk-harnesses ∙ Vercel/geist I'm in awe... such amazing progress and cool libs. Yes - I managed to combine eve with local coding harnesses ~ ftw
8
14
729
Being a contrarian has a secret advantage: You activate your rare almost never used critical thinking brain parts. You try to make sense of why a single post on a new account suddenly spikes on X and Indian news outlets. Ignore the message - think in meta levels Analyze the empty claims. Find research supporting and refuting his claims. And conclude that's hot air abusing fear mongering and psychology of people. My posts went nowhere near the popularity of the doomers because I'm always swimming against the huge current. But I will do so regardless because those rare unused brain regions need some love.
JUST IN: Investigation reveals former Anthropic researcher Jacob Coxon reportedly worked with a “doomer PR firm” while launching his viral campaign warning of AI’s existential risks.
1
11
718
Pausing a subscription so we devs can test and run other models because this thing is moving like hell sounds like an awesome feature request @thsottiaux
Can I bank my unused $200 OpenAI sub for the future reset?
2
9
1,043
Holy shit, the X algo actually inserts people into your "For You" page if you ask for it and matches you?! I always thought this is a myth. Hi @X algorithm. Help me connect with: 💯 Everyone on this platform.
7
14
798
I can't even click "Not interested" in this post. We found our hack.
2
75
What's the goal of these messages? A. Embracing the future that at some point we no longer need to watch code? B. Confidently convincing people not to read code now eventhough they probably should? C. Convincing one-self that the decision to vibecode and ignore all the slop/techdept was okay? D. Clickbaiting some polarizing views just for likes and buy ins for the products? E. Shilling OpenAI & Anthropic just to vendor-lock people into their product? For writing code, I mostly agree. You can let the autocomplete do it's job. If you want to quickly rename a symbol you can still manually do it, it's literally 1 second compared to 30sec to explain it to AI and making sure it doesn't go extreme renaming. What about looking into the code. PRs. Reading what the agent produced. Checking the architecture so it's not "swiss cheese"-like? I certainly welcome scenario A but this is not the reality right now. Fullstop. Esp. because no local model can compete and secure us this position. If you have governments & companies in full control who gets to use this "vibecoding" power and what it only needs is for you to pay half of a rent, then... No thank you. I'm happily using a simple autocomplete and coding by hand for free when I'm offline. And still reviewing my code so it's neatly organized.
It's pencils down, people. Writing code by hand is no longer an economically viable skill for most programmers at most companies. But the future of making software has never been brighter. Don't you dare black pill this beautiful moment! piped.video/vDjW_dRyKXY?si=6Fsf…
5
3
42
1,844
Is it only me or does nobody get any more impression within the first 2+ hrs. My timeline is filled with recent posts and 0 likes for the first 2-6 hrs. Some bangers in between.
5
26
1,239
How is the model performing for you? [1] omg agi — [2] okay — [3] shaddap — [4] stahp May we look into your history regardless of what you just answered? -_-
3
5
694
Reason: nobody measures. It became "the norm". Nah bro, here's $50 capped limit for your Fable/Astra so you can straight up stop AI-first software dev after 2hrs of work. Some people stop, are stuck with local models on an M1 and then are fired because they're "not agentic enough". Others buy their hobby plan (20x) and do literally all their work with that. Result: companies don't notice anything, nobody seems to complain because fire-scared and surprisingly ppl are still productive. Ye, no shit. Until subsidized plan canceling burst the whole AI bubble after IPO. I'll prep the surprised Pikachu meme.
i don't understand why you would limit employee token spend if they are creating value, why would you limit that? so if you limit them, you essentially say: your output is not valuable enough against the token cost but why would 200-500/m be of any value if 5k is not?
2
11
984
Oh no, we're gonna see Opus 5.5 and GPT 6 Sol frontend / games comparisons all day, aren't we?
6
22
909
Nowadays people don't realize it's bad because they no longer read the code.
before ai you would - write some code - it works, ship it - few months later you'd realize it was bad - you'd refactor it - repeat not sure much has changed. you can produce a lot of code now but you can also refactor a lot of code
5
14
1,242