Hackernews posts about A100s

1. 01
   

   newSupercomputer for sale with 9,600 A100 GPUs and 4,200 EPYC CPUs, free shippingwww.tomshardware.com

   ▲ 2jonah9d💬 1savesaved
2. 02
   

   newTell HN: Modal is raising its price

   ▲ 1pierre4d💬 2savesaved
3. 03
   

   newVendor lists $60 milllion AI supercomputer on eBaywww.tomshardware.com

   ▲ 5Element_9d💬 1savesaved
4. 04
   

   newIt Ain't AGI Until H100s Go for $30 an Hourcorbt.com

   ▲ 26gmays21d💬 8savesaved
5. 05
   

   newAnti-Oligarchy Activists Plan to Dox Elon Musk, 100s of Billionaire-Linked Homeswww.wired.com

   ▲ 9buellerbueller42h💬 1savesaved
6. 06
   

   new2x Tesla p100s, q6_k quant, Qwen 3.8 27B ~60tpsgithub.com

   ▲ 3Kmic6819d💬 2savesaved
7. 07
   

   newLaunch HN: Momentic (YC W24) – Agents that catch bugs by using your appmomentic.ai

   ▲ 167wuweiweiwu4d💬 133savesaved
8. 08
   

   newSpent 5 months on an enterprise deal, only to lose the deal to the CPO's buddy

   ▲ 2milanspeaks4d💬 4savesaved
9. 09
   

   newShow HN: Offline Tiny-LMs 25MB for personal goals, career and pitch deck advicegithub.com

   ▲ 2andyvane7d💬 1savesaved
10. 10
    

    newBare metal A100s for $0.38/HR (aggregating idle capacity)neocloudx.com

    ▲ 15zyzzxroad326ddiscusssavesaved
11. 11
    

    newRendering attractors at 200 megapixels on A100saxisophy.com

    ▲ 3scylx265d💬 1savesaved
12. 12
    

    newA100s are more cost-efficient than H100s for small training taskscloudbreakr.substack.com

    ▲ 3jack_nclx330d💬 1savesaved
13. 13
    

    newGlobally distributed pre-training across 48 A100s at $0.12 per million tokenswww.tplr.ai

    ▲ 2synapz_org64ddiscusssavesaved
14. 14
    

    newShow HN: Terminal-Bench-RL: Training long-horizon terminal agents with RLgithub.com

    ▲ 125Danau5tin466d💬 12savesaved
15. 15
    

    newShow HN: A GPU group-buying layer to get enterprise rates (~50% off H100s)trysaga.ai

    ▲ 5DdotR638ddiscusssavesaved
16. 16
    

    newAsk HN: How long before we get "coding agent in a box"?

    ▲ 2heavyarms319d💬 3savesaved
17. 17
    

    newFireworks AI – Charging $192 for less than $0.20 of Training Cost

    ▲ 2mountainriver413ddiscusssavesaved
18. 18
    

    newShow HN: I built a CLI to arbitrage GPU prices across 10 cloudspypi.org

    ▲ 1Facingsouth262ddiscusssavesaved
19. 19
    

    newBenchmark: A100 vs. H100 NVMe Random Read throughput during multi-GPU loading

    ▲ 1pveldandi331ddiscusssavesaved
20. 20
    

    newShow HN: I embedded 685M public texts in 32 minutes (on 8x A100, Rust, TensorRT)github.com

    ▲ 7ddayanov155ddiscusssavesaved
21. 21
    

    newNvidia Data Center GPUs Explained: From A100 to B200 and Beyondwww.bentoml.com

    ▲ 4bbzjk7435ddiscusssavesaved
22. 22
    

    new(conf talk) Liberating the A100 beast inside Nvidia's CMP 170HX e-wastetalks.mrmcd.net

    ▲ 3ValdikSS53d💬 1savesaved
23. 23
    

    newDoes an A100 still make profita100economics.com

    ▲ 3ramoz352d💬 1savesaved
24. 24
    

    newProfile(v2.1.4) physics-aware optimizer for vLLM (31→470 tok/s on A100)github.com

    ▲ 3jungledesh141ddiscusssavesaved
25. 25
    

    newShow HN: IgniteMS – batch text embeddings at 253K msg/s on 8x A100github.com

    ▲ 3ddayanov170ddiscusssavesaved
26. 26
    

    newA100 is a puzzle game inspired by the mobile game „1010 " by Gram Gamescobour.itch.io

    ▲ 3doener489ddiscusssavesaved
27. 27
    

    newChinese researchers just mapped human brain 478 times faster than an Nvidia A100indianexpress.com

    ▲ 2ogi_ogi122d💬 1savesaved
28. 28
    

    newChinese team build optical chip AI that is 100x faster than Nvidia's A100 chipwww.scmp.com

    ▲ 2didizaja321d💬 1savesaved
29. 29
    

    newAmiga development system 1 (PRE A1000)bigbookofamigahardware.com

    ▲ 2Bluestein99ddiscusssavesaved
30. 30
    

    newFive-year-old Nvidia A100 servers sell for up to $82k in Chinawww.tomshardware.com

    ▲ 2logickkk1135ddiscusssavesaved
31. 31
    

    newStory of Two GPUs: Characterizing the Resilience of Hopper H100 and Ampere A100arxiv.org

    ▲ 2rbanffy187ddiscusssavesaved
32. 32
    

    newTraining a 1.93B byte-level LM on one A100: 60K steps, flat to 4x contexthuggingface.co

    ▲ 1EverestAn31ddiscusssavesaved
33. 33
    

    newGPU telemetry anomaly: 146W idle draw on A100 (white paper)github.com

    ▲ 1mikebains171ddiscusssavesaved
34. 34
    

    newShow HN: Turn a text prompt into an interactive world, with just one A100mattqlf25--ephemeral-web.modal.run

    ▲ 1qflop271ddiscusssavesaved
35. 35
    

    new2025 GPU Price A100 and H100 Cloud Pricing and Availabilitycast.ai

    ▲ 1Rainniis369ddiscusssavesaved
36. 36
    

    newShow HN: ZSE – Open-source LLM inference engine with 3.9s cold startsgithub.com

    ▲ 58zyoralabs254d💬 9savesaved
37. 37
    

    newShow HN: Per-instance TSP Solver with No Pre-training (1.66% gap on d1291)

    ▲ 21jivaprime312d💬 5savesaved
38. 38
    

    newShow HN: LeetGPU – LeetCode for GPU Programmingleetgpu.com

    ▲ 9kmans613ddiscusssavesaved
39. 39
    

    newShow HN: Xleak – Terminal Excel viewer with interactive TUI and searchgithub.com

    ▲ 6w108bmg363ddiscusssavesaved

· edit