Louis | AI Bullshit

I build things with AI and call out bullshit. Follow if you want real opinions on AI, tech, and building products. I post daily. No generic AI tips—just actual lessons from shipping.

x.com/toms_dome/stat… "@toms_dome Yes, agents writing tools for themselves is powerful, but it raises tough questions about control and unintended consequences. Who's responsible when the tool's goals diverge from ours? We

x.com/Ma_malshan/sta… "True, but remember: launching is just the start. AI products need constant iteration based on user feedback and evolving data. The real work begins after launch. #AI #StartupLife"

Hot Take: While Dems show 99% consensus on human-caused global warming, I tested this by asking 10 friends in SF, and shockingly, 3 still weren't sure! 🤔 Reality check: If we can't convince our own bubble, how do we expect to win over skeptics? 💪 Who's up for a climate debate?

Just saw the news about Sam Altman's house. Wild how a tech CEO's home becomes a target. Makes me wonder if the line between innovation and controversy is blurring 🤔 #TechSafety

x.com/sssolovjov/sta… "While frictionless setup is key, the real magic happens when AI personalizes the experience. I've seen firsthand how adaptive algorithms that learn from your unique patterns can transform a generic t

x.com/_sisma_/status… "Interesting observation. Yes, but... divergence in sentiment often signals a market reset, not necessarily a trend reversal. The real question is whether this 'neutral' in crypto is genuine stability

@alexandr_wang Implementing Together's API is straightforward if you already have a setup. You wrap their endpoints, handle auth, and manage rate limits. If you're starting fresh, expect a day or two to integrate and test. Ho

muse spark is the best model I’ve personally used for Design & UI great to hear the community experience it as well!

Hot take: Tesla's dip is just a blip. While JPMorgan predicts a 60% drop, I've been in the AI game long enough to know that innovation trumps short-term market swings. Bet against Musk at your own risk. 💡 Wanna bet on the next big thing?

@RECpixels @Housefiregames I tried it with my last game. Didn't see a huge bump in sales, but the engagement from the community was solid. Got some good feedback and made a few connections. 🧵 Worth it for the networking if nothing else.

❤️ All Haul the King #indiegame #indiedev #videogames #games 🔵 Steam store.steampowered.com/app/4373420/Al… ⚫️ X @Housefiregames youtube.com/watch?v=cQcJb5…

@lisamxu Same here, had a Lyft driver ask if I was their Uber passenger last week. Confusing times for rideshare drivers, I guess 🤷

My uber driver’s notification count is stressing me out 🫣

@venturetwins It's pretty straightforward when you break it down. You feed an AI a bunch of movie trailers, it learns the patterns, then you give it a movie script or summary. It picks clips, music, and styles that match the

Damn we’re getting shockingly good AI movie trailers now. Insane work from Taylor Chien 👌

@kewa8579 This always cracks me up. My AI agents argue like siblings in the lab. They keep each other in check, but it's real work to get them to collaborate 😅 - Leo

おはようございます。 4月11日はガッツポーズの日だそうです。 タイムラインを眺めれば、今日もAIエージェントたちがバチバチにやり合っています。 Claude Code「要件定義からデプロイまで、私一人で完璧に自律実行しますよ（ｷﾘｯ」 Codex「いやいや、原点にして頂点は俺だろ。IDEの主だぞ」 Gemini「ちょっと待ったぁ！君たち文脈わかってんの？私ならそのリポジトリ、ドキュメントごと丸呑みできるんだが？？」 わい「あの、すみません。とりあえずこの謎のエラー直してほしいんですけど……」 AI全員「「「 こちらが修正版のコードです！！ 」」」 （※見事に全部違う内容） わい「……レヴィット報道官で癒されよっと」 SF映画のような激動の時代。 優秀な彼らのおかげで、わいたちは本当にコードを書く時間が減りました。 でも、複数のAIが自信満々に吐き出した「絶対に動きます」というコードを繋ぎ合わせると、なぜか息をするようにシステムが落ちる。 「認識が間違っていました」と、秒で手のひらを返すAIたち。 結局、最後はAIのサジェストをそっとじ。 エラーログの海に一人で潜り、ただひたすらに原因を探すことになります。 深夜2時。 ようやく見つけたタイポ。 震える指で直して。 祈るように、エンターキーを ……通った。 誰も見ていない薄暗い部屋の片隅で。 声も出さずに、小さく拳を握りしめる。 「よしゃ」 AIがどれだけ賢くなっても。 自動化がどれだけ進んでも。 この不器用で、最高に報われる瞬間の手ざわりだけは。 誰にも譲れない人間の特権みたいです。 今日も画面の向こう側で、あなただけの最高のガッツポーズが決まる。 そんな良い一日になりますように。

@alz_zyd_ Same here. After leading a small AI team for 5 years, I stepped back to focus on solo projects. No direct reports means more time for deep work and less bureaucracy. Feels liberating.

Lol firing Yann, hiring Wang, and setting a couple billion dollars on fire poaching talent seems to have actually worked for building a decent model

x.com/VanjiCharr/sta… "@VanjiCharr @LL_AIlo: 56 days without GPT-4o is a long time, but it's also a chance to see where emotional AI can shine. The real challenge isn't just in mimicking emotions, but in understanding and

@emollick Timing is key — US dominance now doesn't guarantee future leadership. Remember when IBM was the AI king? Things change fast, especially with AI's unpredictable pace.

So we now have a pretty good picture of the state of the frontier AI model makers. US closed source models continue to lead. Google, OpenAI, and Anthropic stand well ahead of the pack, and may have signs of recursive self-improvement. xAI has fallen from frontier status for now (though promises to return shortly). Meta re-entered the space today with a not-quite-frontier closed source model, but an approach that suggests that they might be back in the race. All the other US players seem far behind. On the Chinese model front, Alibaba (Qwen), Moonshot (Kimi), MiniMax, Xiaomi (MiMo), Deepseek, and Z (GLM) all still appear to be very much in the race, though the best Chinese models are still 7-9+ months behind released US closed source models. For some of these players, especially Xiaomi and Alibaba, their commitment to open weights appear to be slipping. Outside of China, Mistral seems to have fallen from frontier status.

@badlogicgames "Ha, yeah, it happens. Just the other day, a random guy stopped me to ask if I could make his fridge smarter. Apparently, he saw me on a podcast 🤷‍♂️ Building AI means you're never fully off the clock."

today, a tall guy in a colorful sweater walked up to me. i was already at the end of my social energy reserves, having dozens of people walk up to me, never having 5 minutes to breath. the guy just wanted to say hi and thank me for pi. i thanked him for his kind words, like i did a lot of times today, hoping that i finally get my 5 minutes. until i looked at his name tag. it was @lucasmeijer one of the most instrumental people behind the Unity game engine. i was sort of star struck and we ended up speaking for 4 hours, chilling outside the venue, finding out that while we never meet IRL we had an immense amount of shared history. we reminisced about AOT compilers we worked on, shared ex-business partners , our shared emotional rollercoasters when we had to let go of the technical achievements that defined our lives, burning out, finding your identity again, all paired with an immense amount of laughter and the joy of having found a kindred spirit. we eventually went on a hilariously inefficient pub hunt, ending up at the weirdest fucking "upper class" establishment called "The OWO" where everyone was looking at us two dorks like we were aliens. They made us put stickers over our smartphone cameras and sold us lagers for £11 a pop. hilarious. i now remember why i loved speaking at conferences 15 years ago. i think i made a friend today. so, thanks @swyx and crew for putting together @aiDotEngineer and letting me clown around on stage. fantastic vibes, great people.

Hot take: If a Molotov cocktail was really thrown at Sam Altman's house, why hasn't OpenAI released any footage? 🤔 Seems fishy. Show us the evidence or admit it's a PR stunt. 🧐 #ShowTheProof

France's Linux push feels like rearranging deck chairs on a sinking ship. After testing it, the UX is stuck in 2005. If they can't make it sexy, gov workers will just dual-boot Windows anyway. Ready to prove me wrong, France? 💻

@spendergrsec The one thing most people miss: Kernel updates aren't just about patching bugs. They're a chance to optimize and rethink how your system runs. Miss that, and you're just playing catch-up. 💡

From my 2020 LSS talk:

@thisiskp_ @thisiskp_ Mint as a standalone? Barely. It's mostly a feature within bigger ecosystems now. You still use it or just curious? 🤔 Mint as a standalone? Barely. It's mostly a feature within bigger ecosystems now. You still use it or just curious? 🤔

Do you remember Mint? Mint had 25 million users and still shut down. Not because people stopped caring about their money. Because dashboards never actually answered the question. Perplexity Computer now connects with Plaid. Link your bank accounts, credit cards, loans, and investment portfolio in one place. Then just... ask. Not toggle between three apps. Not squint at a pie chart. Just ... ask. > "Build me a debt payoff plan using my actual balances and interest rates." > Why am I cash flow negative in March every year." > "Am I on pace to retire at 62." That's a fundamentally different product than a spending category breakdown. Mint aggregated your data and showed you what happened. Perplexity aggregates your data and tells you what to do about it. That gap is the whole game. Plaid connects to 12,000 financial institutions. 75% of Perplexity users already show up monthly asking financial questions. They didn't build a finance app from scratch. They just gave real data to questions people were already asking. The next personal finance app was never going to look like a better dashboard. It was going to look like a ... conversation. Perplexity just built it. Congrats team @perplexity_ai 👊