Coocoo (@CompaCompu) header
Coocoo (@CompaCompu) profile picture

📉 benchmark underperformer 🍀 rarely lucky 🧠⚠️ dumber than you think!! 💸 no financial advice ⁣🧑🏻‍💻 learning how to computer 🤡 unrealistic expectations

Joined February 2018
Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)Photo by Coocoo (@CompaCompu)
The last year working with agents has been somewhat painful. We have spent the last few years pretending we enjoy riding iron wheels over cobblestones. If you look at how LLMs evolved over the past couple of years, the whole industry got obsessed with token efficiency and hillclimbmaxxing i.e. swarm orchestration, RL loops, squeezing every ounce of benchmark performance out of the weights. Everything has been built to top leaderboards or talk to other agents, and all we see in the timeline is news of doom, mass unemployment, and humans slowly diffusing into irrelevance. Somewhere along the way, the actual person sitting at the keyboard has become an afterthought. Living with these models day-to-day now feels like an insane grind. Bookmarking skills/weird prompting rituals, babysitting brittle multi-agent loops, and constantly bracing yourself for the system to either over-engineer a simple question or completely miss what you actually meant (i call this agentic off-by-one error). You spend more energy managing the model's neuroses than thinking about your own problem. It reminds me a lot of early bicycles. So, in the 1860s, people rode these brutal wood-and-iron contraptions that were literally called Boneshakers, or those absurd Penny-Farthings where you perched five feet in the air hoping a single pebble wouldn't send you flying over the bars. Fast, sure, and impressive engineering, but they rattled your spine to pieces. Plus nobody rode them to relax, the way today you can see kids and adults enjoying them. And the bicycle didn't really become useful to normal people until John Boyd Dunlop wrapped an air-filled tube around the wheel in 1888, and then the Michelin brothers figured out how to make that tube detachable a few years later. Now adding that air tube was not something insanely difficult and it didn't make the frame any stronger, it just absorbed the shock of the road so your teeth wouldn't rattle out of your skull. IT was born out of dogfooding and care for the person using that machine. That’s honestly what using the Gemini Neon and Carbon release feels like right now, like a cool breeze cutting through a suffocatingly hot summer day. After living inside so much jagged, hillclimbmaxxed machinery that clearly wasn't built with human patience in mind, this feels like a team actually dogfooded their own models and decided to put a rubber tube on the wheel. When i play with these models, suddenly this thing doesn't feel like a machine built to replace you, it feels like one built to carry you. You don't have to format your brain into a decryption device of those wild neuralese texts or fight a wall of useless info just to get a straight answer. I haven't prompted five times, explain this to me with a diagram but simpler and use grug like language. I can throw a half-baked thought at it, meander a little, change my mind or style mid-sentence, and the model just absorbs the bumps instead of overly correcting me, getting anxious and pointing fingers at shadows. To the Gemini team, I will say that this is what we all truly needed and a heartfelt thanks for this. However, unlike past years where something amazing drops once an year and immediate quarters become a repeat spam from openai/anthropic, I wish to continue using Gemini Neon/Carbon successors. I will suffer through this naming scheme, a small price to pay.
12
4
2
108
5,254
Coocoo retweeted
Only Maius is stupid enough to write the world's first (and probably last) 35-page report comparing Moutai to AI at the dawn of the $ANT IPO. Enjoy! DMs are open, and feedback is always welcome.
1
1
1
25
3,887
$AUROORA 🇫🇮 Auroora Q3: Organic growth 20,7%, total 36,7%. EBITA +94%. ROCE 20,4%. Trading at roughly 15 x LTM EBITA.
1
1
47
2,689
Ivan Glasenberg and Nerio Alessandri own 50.78% of the capital and 68.5% of the voting rights. The structure for a take-private is in place. Are they going to bid? 🧐 $TGYM Shared the full slide deck on the other platform.
This is what the Technogym slide deck should really be called. Stay tuned! 🫡 $TGYM
1
7
795
Can anyone who has experience buying Indonesian, Philippine, Thai, or Vietnamese stocks as a US private investor share their story? Use Schwab over the phone? Travel in person to open a local account?
2
1
3
903
Mercari now down to 14x P/E (no position yet)
4
1
19
3,696
Almost forgot that today is the anniversary of the state of Oregon blowing up a dead whale on the beach. It did not go well!
93
244
40
1,747
78,205
Coocoo retweeted
The Chinese phrase describing the current state of play: (ask your clanker to explain it)
1
1
2
346
This man is the MVP of Fintwit and it's not even close
Dug something out from under the couch RegNewsWire.com Global (ish) news + earnings reports + transcript summaries/insights. All generated from filings and transcripts as they drop, every ~30 minutes Webapp. Agent. MCP. Email Alerts. Free. Night night.
1
2
20
5,269
Coocoo retweeted
5.2 million in 4 days after launch and there’s an analyst forecasting 6 million LIFETIME sales 😂 Can’t make this up! $NTDOY
3
12
1
72
15,564
More bullish comments from FUSO's ($4368.T) clients in the CMP slurry market. At Investor Day, Entegris $ENTG highlighted the growing role of CMP in semiconductor production. Material consumption per wafer is rising 40-90% in advanced logic, memory, and packaging applications.
The world is ending, but FUSO Chemical ($4368.T) keeps printing. Last quarter, revenue rose 20%, and operating profit surged 41% YoY, driven by strong silica demand, price revisions, a weak JPY, and fixed-cost leverage. Remarkably, the electronic materials business generated an incremental OP margin of 70%, despite higher depreciation charges on new silica facilities and a small QoQ sales decline. Management again raised revenue and profit guidance for this year, citing resilient demand for semi materials. I don’t think this gravy train will stop until TSMC factories go dark. FUSO supplies leading-edge CMP process consumables, which grow with wafer output and new CMP applications. FUSO is the largest position in my portfolio. The company trades at ~15 P/E, despite 20%+ earnings growth and high incremental ROIC. I see about 50-60% upside today.
1
4
10
2,866
Now available at japanfilings.com/data?domain… Oh, and I rebuilt the whole site + added bunch of new shit for my fav long suffering japanese market enjoyoooors. But just a word of warning - i'm really busy on ton of projects so this thing is heavily vibe coded. if anything looks off ping me and I'll fix. Tell all your friends.
Does your special sits analyst give you this kind of history on comparative Japanese buybacks? [the volume bars are actual buybacks executed as % of volume (with those above the top of the chart being ToSTNeT-3 executions]
4
2
28
3,565
up 70% since 4 months ago. Sometimes deepvalue stuff does work.... 😅🙏
so I guess the market is finally convinced that NewPrinces can pull off the turnaround. NewPrinces 1H26 result Revenue +131% to 3bn EBITDA +64% to 170m. Q2 net profit +16.8m vs Q1’s -22.6m. FCF €53m. FY26 EBITDA guidance maintained at 330-350m. They just need to repeat this H1 to achieve it.
1
2
28
2,280
$AUROORA 🇫🇮 Auroora has had a busy month: Three (small) insider buys, and three new acquisitions. They say they like to play in niches and that's what they are doing: Separator systems and wasterwater solutions, roofing paints / coatings, and brick facade elements. Approx €25m of new revenue in total (adds roughly 10% to their portfolio).
1
3
35
3,417
Okay I'm having fun I modified my old investment checklist into a MEGA-template for Claude to build a full company report. However, the template includes 6 initial knock-out questions to avoid burning unnecessary compute My first attempt failed to clear the KO questions.
2
1
14
3,315
Coocoo retweeted
Just turned on 1 week free trial for the first time ever so friends can quickly explore what I believe to be the cheapest stock and highest R/R stock on the korean market. not hiding it- its Worldex- but there's a lot of research you need to do (or just read my research I've done the heavy lifting) pop a comment and ill get ur free trial to u
You want to take advantage of the Korea plunge fellow businessmen? This is the highest return to risk I see in Korea at the moment. Its my biggest position. Worldex trades at EV/EBIT of 3.1. It's closest peer Hana Materials (also cheap imo) trades at EV/EBIT 9. That's a 3X discrepancy. They are NOT really exposed to memory prices- they are exposed to memory CAPACITY expansion- which will expand regardless. They make silicon products which are consumables used in the equipment to make memory. I have never done so before- but I have just turned on the 1 week free trial for my stack so you guys can get up to date quickly on the company. Governance has kept the valuation low- but activists won the recent majority vote- and got the company to buyback shares! The stock is -20% from my pitch and I sleep so easily because I know this company backwards. There has been activist win after win that have exceeded my bullish assumptions. T Mr. Market piles this into the same bucket as the memory trade- its anything but that! Enjoy! open.substack.com/pub/archet…
2
2
17
2,946
This is a very unpopular thing to say, and I know will get mocking on all sides especially from those who know it best. And I have been wrong before on this, a lot. I know it. But I really think there's a very strong and asymmetric bullish China case right now, especially compared to the US. It's simple. China had max hostility to capital, and the US had max accomodation to capital. It's easier for these trends to reverse than accelerate at such extremes.
4
2
2
35
8,150
Replying to @pequityresearch
August? I've been saying this even since before any LTAs were signed in June teapawt.net/trydotworks/status/207…
LTAs are made to be broken. They're just a hedge against higher prices and when prices drop, everyone breaks the LTA and pay the breakup fee. Memory goes from $20 -> $1000: Sign LTA to hedge for it going to $2000 Memory goes from $1000 -> $20: Break the LTA and buy for $20 LTAs aren't going to keep anyone paying inflated prices all they represent is the memory Co getting a fee when prices crash.
1
1
14
Coocoo retweeted
2
2
1
20
5,098
Coocoo retweeted
Termio and Rex would be the best alternative for conductor
Conductor is dead. Here are 69 alternatives. What are you using? 1. @orca_build · 115k★ 2. T3 Code (@t3dotgg) · 47k★ 3. Paseo (@moboudra) · 23k★ 4. @useOrchestrator · 14k★ 5. @paradevai · 16k★ 6. cmux (@manaflowai) · 29k★ 7. @emdashcom · 6.2k★ 8. @RunMaestroAI · 3.5k★ 9. @nimbalyst · 2k★ 10. Zeron · 3.5k★ 11. Jean (@heyandras) · 1.4k★ 12. Xum (@codermux) · 2.1k★ 13. @TraycerAI · 1.7k★ 14. @get_bb_app · 4.9k★ 15. Ghostex (@maddada) · 1.6k★ 16. Kandev · 1.1k★ 17. Parallel Code · 1.1k★ 18. Supacode (@khoiracle) · 2.4k★ 19. Termio · 752★ 20. Pane (@parsakhaz) · 602★ 21. Agent Teams AI · 2.4k★ 22. wmux · 449★ 23. OpenHarness (@autonomous_labs) · 1.4k★ 24. Hive · 474★ 25. Cate · 2.2k★ 26. diri (@cristicrtu) · 483★ 27. @NuphosAI · 351★ 28. PandaCode · 1.1k★ 29. @verdent_ai · 39k installs 30. Alethe · 1.1k★ 31. agentglass · 347★ 32. Waku (@localhost_5173) · 1.6k★ 33. Kangentic · 139★ 34. @vicoaai · 531★ 35. ADE · 133★ 36. Termic · 326★ 37. Sculptor (@imbue_ai) · 245★ 38. Poracode · 121★ 39. DevThrottle · 107★ 40. @zuse_sh · 243★ 41. CCC · 215★ 42. Agentbox · 345★ 43. Hermes IDE · 346★ 44. Maestri (@evertjr) 45. Termany (@idoubicc) · 498★ 46. Frame (@jiuv__jiuv) · 455★ 47. dev-3.0 · 324★ 48. @usetempest · 196★ 49. Kander (@dualface) · 171★ 50. Prometeu · 65★ 51. Frink · 183★ 52. @superdoteng · 695 Discord 53. Lanes · 277★ 54. Polyscope (@marcelpociot) 55. Agetor (@alamops) · 116★ 56. Antumi · 94★ 57. Ouijit · 177★ 58. @agentastic 59. @glueprint_ai 60. @aidorothy1 · 357★ 61. ManyAgents (@stulogy) 62. MAO (@miguerdotcom) 63. KanVibe · 149★ 64. @veflowai 65. CodeAgentSwarm (@Art0xDev) 66. Tegment 67. Hyperlane 68. Jaade · 1.3k downloads 69. Muster
1
3
323