ice.computer
→
ML News
refresh
1.
▲
Running local models is good now
1530 points
(vickiboykis.com)
1 day ago
2.
▲
GLM-5.2 is the new leading open weights model on Artificial Analysis
845 points
(artificialanalysis.ai)
23 hours ago
3.
▲
Sixty percent of US consumers say 'AI' in brand messaging is a turnoff
1040 points
(wpvip.com)
20 hours ago
4.
▲
Feds freaked over Fable 5 after 'fix this code', not jailbreak, say researchers
600 points
(www.theregister.com)
1 day ago
5.
▲
Midjourney Medical
627 points
(www.midjourney.com)
6 hours ago
6.
▲
Leaked financial docs show OpenAI is losing billions of dollars a year
353 points
(arstechnica.com)
10 hours ago
7.
▲
Has AI already killed self-help nonfiction books?
399 points
(tim.blog)
1 day ago
8.
▲
Only 16 Percent of Americans Think AI Will Have a Positive Impact on Society
379 points
(techcrunch.com)
15 hours ago
9.
▲
AI demands more engineering discipline. Not less
377 points
(charitydotwtf.substack.com)
18 hours ago
10.
▲
A robot is sprinting towards you. Do you want it running on Claude or Grok?
240 points
(openrouter.ai)
11 hours ago
11.
▲
GPT‑NL: a sovereign language model for the Netherlands
249 points
(www.tno.nl)
1 day ago
12.
▲
The founder's playbook: Building an AI-native startup
228 points
(claude.com)
1 day ago
13.
▲
Qwen-Robot Suite: A Foundation Model Suite for Physical World Intelligence
210 points
(qwen.ai)
1 day ago
14.
▲
Wolfram Language and Mathematica version 15
218 points
(writings.stephenwolfram.com)
1 day ago
15.
▲
Leaked OpenAI financials show $38.5B loss and compute burn
220 points
(runtimewire.com)
1 day ago
16.
▲
Claude: Elevated errors across many models [resolved]
189 points
(status.claude.com)
1 day ago
17.
▲
Launch HN: Adam (YC W25) – Open-Source AI CAD
180 points
(github.com)
16 hours ago
18.
▲
Local Qwen isn't a worse Opus, it's a different tool
164 points
(blog.alexellis.io)
5 hours ago
19.
▲
AI is potentially a Dunning-Kruger effect amplifier
66 points
(twitter.com)
1 day ago
20.
▲
Show HN: High-Res Neural Cellular Automata
193 points
(cells2pixels.github.io)
23 hours ago
21.
▲
Using AI to improve a challenging reaction in medicinal chemistry
53 points
(openai.com)
14 hours ago
22.
▲
Show HN: VoiceDraw – Talk system design out loud, the diagrams draw themselves
47 points
(voicedraw.com)
1 day ago
23.
▲
DeepSeek V4 Pro at 5% the cost of Claude – what it takes to close the gap
54 points
(howardchen.substack.com)
1 day ago
24.
▲
Show HN: I built 184 free browser tools – PDF, image, dev, AI tasks, no upload
87 points
(brevio.pro)
22 hours ago
25.
▲
SubQ 1.1 Small::
129 points
(subq.ai)
1 day ago
26.
▲
The octopus architecture for AI agents
22 points
(blog.goodman.dev)
1 day ago
27.
▲
Meta head of product for 'AI for work' transformation is leaving company
22 points
(www.reuters.com)
5 hours ago
28.
▲
Apple Intelligence may become mandatory in iOS and macOS 27
24 points
(manualdousuario.net)
12 hours ago
29.
▲
France to ditch Palantir's AI data tools in favour of domestic provider
39 points
(www.theguardian.com)
1 day ago
30.
▲
AI Compute Extensions (ACE) Specification
32 points
(x86ecosystem.org)
5 hours ago
31.
▲
DeepSeek Introduces Vision
77 points
(chat.deepseek.com)
2 hours ago
32.
▲
Multiple JetBrains IDE plugins caught stealing AI keys
30 points
(www.aikido.dev)
23 hours ago
33.
▲
GateGPT: 56k tokens per second Transformer (KV cache) on FPGA at 80 MHz
40 points
(twitter.com)
1 day ago
34.
▲
Show HN: Tyto – find where audio breaks your voice-agent calls
15 points
(call-analysis.ai-coustics.com)
17 hours ago
35.
▲
AI will lead to labor shortages, Bezos says in optimistic talk
15 points
(www.reuters.com)
16 hours ago
36.
▲
Pentagon boasts of using AI to write reports mandated by Congress (1.5mil users)
73 points
(arstechnica.com)
18 hours ago
37.
▲
GLM 5.2. Performance Benchmarks
147 points
(artificialanalysis.ai)
1 day ago
38.
▲
GLM 5.2.2
16 points
(huggingface.co)
1 day ago
39.
▲
GLM-5.2: Built for Long-Horizon Tasks
30 points
(z.ai)
1 day ago
40.
▲
GLM-5.2: Frontier Intelligence, Open Weights
34 points
(twitter.com)
1 day ago
41.
▲
GLM-5.2: Built for Long-Horizon Tasks::
30 points
(z.ai)
1 day ago