Marcus Eisele (@eiselems) header
Marcus Eisele (@eiselems) profile picture

Software engineer, 20y+ years. Exploring AI in real work and everyday life. Coding agents, local AI on a 3070, and experiments worth sharing.

Stuttgart Area, Germany
Joined August 2015
Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)Photo by Marcus Eisele (@eiselems)
I'm not quitting my 9-5. I'm building toward having that choice. Exploring AI agents, local AI on my RTX 3070, and ideas that could earn on the side. Sharing what works, what fails, and what I ship. Anyone else figuring this out alongside a full-time job?
4
557
Never gave it a second thought, GTA IV will be the peak of "almost only human exclusively made" Games. Was not aware that game is such a general milestone besides its general hype
Being upset that any sw - game, app, a website etc - uses AI for development will soon be as sensible as being upset that any of these things use a compiler to produce code. Majority of software produced today is already with AI. Last major non-AI-assisted game will be GTA VI
117
For sure the next family gathering will be interesting ...
If anyone doubts the power of Chinese pussy, show them this
1
104
Going back to other models after using Opus kind of sucks, ngl
1
101
What are y'all doing to hit token limits on multiple codex/claude plans? genuiely curious
3
6
426
At this point I don't care what people think :D Who wouldn't say no to this card?
We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄 To enter: 🟢 Share this post 🟢 Comment #RTXON Terms and Conditions: nvidia.com/en-us/geforce/con…
1
98
Marcus Eisele retweeted
We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄 To enter: 🟢 Share this post 🟢 Comment #RTXON Terms and Conditions: nvidia.com/en-us/geforce/con…
84,949
72,382
7,718
67,796
5,968,091
Qwen release is a major indicator for progress of open weight models, so pumped for this Amazing how relevant Qwen 3.8 27b still is
The becoming begins.
1
88
Asked Claude what symptoms of Ai psychosis are... One of them was "Ideas of reference: everyday details such as license plates, timestamps, or coincidences start to feel like signs." If I see this I can reach out to the "Hotline 0800 111 0 111"... Only binary 😰
117
Quite amazed what I vibecoded for my kids still somehow can't be assed to translate it to english just for posting it here 🤣
103
Seeing such a benchmark and the best model getting 13%, I often question myself, how would the world change when a models starts saturating that benchmark? What do you think a world would look like if we have models that start hitting 90% on that one? Any wild takes?
Today we’re releasing GDPx, a benchmark for measuring agentic capabilities across the most valuable domains. GDPx is a true frontier-level benchmark: even the best model scores just 12% (1/7)🧵
2
228
Gave my agent everything it needs for generating a graphical game for the kids ... let's see what it will be cooking
1
180
I would say why don't we only give food to restaurants ...
WTH: Sam Altman complaining that individuals (like me!) are purchasing RAM for our PCs wanting to run local inference (and do other stuff), because it gets in the way of him buying more RAM to sell inference at 80% margins… this is barking up the wrong tree
1
169
Weekend project: neuhausen-detektiv.marcus-ei… Started building a PoC for a point and click adventure within our hometown Tech used: • Qwen 2.1 Image Turbo (based on reference Image) • Qwen3-TTS-12Hz-1.7B-VoiceDesign • Opus 5.5 driving the whole development Curious about more?
1
3
228
Thinking 5 to 10 years back I never would have considered it. For me I always was a windows/Linux for the stationary and Mac book for being on the go. Today I would really consider a Mac studio for the massive unified memory and a Linux laptop just due the easy setup with agents if haptic keeps up 🤣
I opened the box form @DHH and it's the Dell XPS 14 with @omarchy loaded on it! Set up was very easy, fast and flawless, I think it took less than a minute on first opening it Omarchy asks you lots of questions to set it up for you, kind of like an onboarding, but not in an annoying way like Windows would do these days with too many useless questions, every question here makes sense It sets up your Git and your AI agent of choice like Codex, Claude Code or OpenCode Everything worked flawlessly with no bugs at all, really smooth Then you're inside, and you see the famous Omarchy wallpaper, and then it gets a bit hard (for me at least) because I had to figure out the shortcuts Luckily @jankeesvw helped me figure them out but I'm still learning I'll be setting up my SSH to my VPS servers so I can work from there. And also I'd like to see if I can use Claude Code to customize Omarchy to my own liking, like generating my own apps etc. I'll report back! Part 2 of the video is in the reply (after I fixed the power)
146
Ouf that hurts, at this point you really must ask yourself if it is worth to go from 256 to 512gb 😅
2
5
358
Received a steam deck from my brother "before it collects dust" Not getting warm with it for now. I added a global voice to text button which uses whisper for voice input most likely using that for chatting without a controller. Any more nice suggestions?
1
124
This is more often what I care about, sadly my dollars are not on a logarithmic scale when I pay 🤣
When using LLMs there are two kind or people, people using a plan or people using the API. The Artificial intelligence index had a logarithmic scale. Sometimes I wonder why. If you pay API prices this chert might be for you
2
2
282
When using LLMs there are two kind or people, people using a plan or people using the API. The Artificial intelligence index had a logarithmic scale. Sometimes I wonder why. If you pay API prices this chert might be for you
1
1
348
Imho it comes down to speed and cost per task, wouldn't mind something faster and a bit more clever than Luna
GPT-6.1 Terra: delivering fast and capable computer use at a lower price. GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.
1
1
196
Next fable release will drive people into AI psychosis if it is a big step up compared to opus
Opus 5.5 is one of anthropics biggest success. on OpenRouter token volume went up 3.6x since launch. And I havent heard anyone saying Opus 5.5 is a bad model. Its the oppostite. Everyone loves it. If they stick to this approach, Fable 5.5 will be the biggest success ever. No wonder they keep it until IPO.
100