2,244 karma · joined April 6, 2026
Yes, we have all seen the math theorems being proven... just higher processing power at the service of the same algorithmic and conceptual patterns? [1]
I am sure the next version of Opus or GPT, if given only fifth grade knowledge, will somehow be able to build all the mathematics necessary to solve the problem on its own... right? Right?
[1] - "AI Isn’t Outthinking Mathematicians. It’s Out-Remembering Them" - https://davidepiffer.com/p/ai-isnt-outthinking-mathematician...
And for the inevitable critics of Sabine...maybe Leonard Susskind is good enough for you: https://youtu.be/2p_Hlm6aCok
They though they would laugh all the way to the bank...but that laugh was already the problem...
"We Need To Talk About Leopold" - https://youtu.be/rE75WvOtcu8
Most of these are 2026....
Frontier LLMs Still Struggle with Simple Reasoning Tasks - https://arxiv.org/abs/2507.07313
General365: Benchmarking General Reasoning in Large Language Models Across Diverse and Challenging Tasks - https://arxiv.org/abs/2604.11778
LLMEval-Logic: A Solver-Verified Chinese Benchmark for Logical Reasoning of LLMs with Adversarial Hardening - https://arxiv.org/abs/2605.19597
LogicGraph: Benchmarking Multi-Path Logical Reasoning via Neuro-Symbolic Generation and Verification - https://arxiv.org/abs/2602.21044
Blind-Spots-Bench: Evaluating Blind Spots in Multimodal Models - https://arxiv.org/abs/2607.08317
Vision-Language Models Lag Human Performance on Physical Dynamics and Intent Reasoning - https://arxiv.org/abs/2601.01547
Do Vision-Language Models Understand 3D Scenes or Just Catalogue Objects? - https://arxiv.org/abs/2605.20448
The Reversal Curse: LLMs Trained on “A is B” Fail to Learn “B is A” - https://arxiv.org/abs/2309.12288
Large Language Model Reasoning Failures - https://arxiv.org/abs/2602.06176
"General365: Benchmarking General Reasoning in Large Language Models Across Diverse and Challenging Tasks" https://arxiv.org/abs/2604.11778
"...General365, a benchmark specifically designed to assess general reasoning in LLMs. By restricting background knowledge to a K-12 level, General365 explicitly decouples reasoning from specialized expertise. The benchmark comprises 365 seed problems and 1,095 variant problems across eight categories, ensuring both high difficulty and diversity. Evaluations across 26 leading LLMs reveal that even the top-performing model achieves only 62.8% accuracy, in stark contrast to the near-perfect performances of LLMs in math and physics benchmarks..."
"Frontier LLMs Still Struggle with Simple Reasoning Tasks" https://arxiv.org/abs/2507.07313
"General365: Benchmarking General Reasoning in Large Language Models Across Diverse and Challenging Tasks" https://arxiv.org/abs/2604.11778
"...General365, a benchmark specifically designed to assess general reasoning in LLMs. By restricting background knowledge to a K-12 level, General365 explicitly decouples reasoning from specialized expertise. The benchmark comprises 365 seed problems and 1,095 variant problems across eight categories, ensuring both high difficulty and diversity. Evaluations across 26 leading LLMs reveal that even the top-performing model achieves only 62.8% accuracy, in stark contrast to the near-perfect performances of LLMs in math and physics benchmarks..."
So what happen is a prompt said for example, find freelance opportunities in Ireland but keep in mind some of these might be available via temp agencies in London.
If you offer me not freelance but permanent roles, and not in Ireland in London...that is a logic failure.
Its this type of complexity with the normal world, that these SOTA constructions so badly fail at, and so spectacularly fail at the margins... despite maxing all benchmarks...Parlour trick.
I must be taking crazy pills and the AGI surely will pass me by... But TODAY, middle August 2026...And in the context of testing and evaluating the capabilities of current SOTA models to implement an Agentic application for job search, here is some simple inhouse built evals I run today, since I don´t trust LLM vendors published benchmarks...
Models tested: GPT-5.6 Sol in Extra High mode and Opus 4.8 Max.
TASK REQUEST: Clear, not too long not too short prompt, for LLMs to go out and research freelance consulting gigs for one specific IT domain, and in one specific country in Europe, including maybe opportunities driven from temp agencies based in geographically close countries.
RESULT: Models go out, fetch the data, and completely misunderstand the task...offering on first results, permanent roles instead of freelance, and based on the country where the agencies are, not in the one it was request for. Think for example IT jobs in Ireland, while freelance agency in London.
ANALYSIS: No intelligence I can call it shown by models, adding cognitive effort for human in the loop to detect subtle factors, and therefore totally useless for agentic app...Best practices would be I guess to add agents on top of agents but although in the p95 of cases that will reduce the errors...for the remaining 5% that could have hallucinations or logic hallucinations like these ones, compounding on top of other logic hallucinations.
I dont care about the theorems being proven. At the end we will found out what most mathematicians were doing, was just exploring the same combinatorial and abstraction patterns. And because of that I am sure LLMs will make mince meat of a lot of mathematical domains.
But right now, what we call intelligence is not existing where it matters, and Ed Zitron is right its a parlour trick.
You mean like NVIDIA is doing to their GPUs for Crypto?
The scientific basis for this statement is unclear.
Well in this case, for the idea it was already in the training data. User said was out of ideas...trying to prop up the LLM smartness...
https://www.lifewire.com/how-to-find-a-lost-bluetooth-device...
https://android.stackexchange.com/questions/194584/locating-...
https://forums.macrumors.com/threads/i-lost-my-cellphone-som...
https://discussions.apple.com/thread/3317786?
And about the code...well...
- Complete Android code example specifically explains that Bluetooth RSSI can be used, cautiously, as a distance indicator. - https://mobile.fhstp.ac.at/development/rssi-messungen-von-bl...
- android-Bluetooth-RSSI-Scanner - https://github.com/TimingJL/android-Bluetooth-RSSI-Scanner
- https://stackoverflow.com/questions/22914532/ble-continuous-...
- https://github.com/alt236/Bluetooth-LE-Library---Android
These LLMs are mainly knowledge bases and compression engines.
05:05 - What are your reactions to Big Tech earnings? - https://youtu.be/vE_FR0O-Jhk?t=306
06:19 - Does the negative free cash flow concern you? - https://youtu.be/vE_FR0O-Jhk?t=380
07:53 - What type of companies were these five years ago and what type of companies are they now? - https://youtu.be/vE_FR0O-Jhk?t=474
09:30 - Do you agree with the markets right now that the CapEx has gotten out of control? - https://youtu.be/vE_FR0O-Jhk?t=571
12:21 - What are your thoughts on Apple making a big bet and deciding not to engage in the AI CapEx wars? - https://youtu.be/vE_FR0O-Jhk?t=751
15:29 - Is it possible the hyperscalers still don’t know the ROI on their AI investments? - https://youtu.be/vE_FR0O-Jhk?t=931
17:43 - How big of a problem is it if all of that is true? - https://youtu.be/vE_FR0O-Jhk?t=1065
22:53 - Are these companies trying to give themselves corporate Botox to look young again? - https://youtu.be/vE_FR0O-Jhk?t=1375
28:13 - How should investors value companies that are deeply tied to AI frontier labs? - https://youtu.be/vE_FR0O-Jhk?t=1697
36:07 - What do you make of the shift from a supply crisis to a potential demand crisis? - https://youtu.be/vE_FR0O-Jhk?t=2169
38:33 - Am I overstating the threat of these inexpensive LLMs out of China? - https://youtu.be/vE_FR0O-Jhk?t=2314
44:05 - Do you think that OpenAI and Anthropic will ever figure out the unit economics? - https://youtu.be/vE_FR0O-Jhk?t=2646
47:23 - How much of this anxiety is priced in right now? - https://youtu.be/vE_FR0O-Jhk?t=2844
49:38 - Are you comfortable with current valuations in Big Tech? - https://youtu.be/vE_FR0O-Jhk?t=2979
51:51 - Are companies using their credibility to reduce shareholder risk or hide their debt? - https://youtu.be/vE_FR0O-Jhk?t=3112
54:45 - Would you ever value OpenAI or Anthropic? - https://youtu.be/vE_FR0O-Jhk?t=3286
56:15 - Would you ever value OpenAI without the full story? - https://youtu.be/vE_FR0O-Jhk?t=3378
58:59 - What are your reactions to SpaceX? - https://youtu.be/vE_FR0O-Jhk?t=3540
01:00:45 - Where do you see the parallels or not from 1999? - https://youtu.be/vE_FR0O-Jhk?t=3649
The analogy I like is that building software is running a Michelin restaurant. The moment you scale, the chef is just writing cooking books and is absent, and you move into franchising, you will be amazed at the bottom line revenue scaling, while customers will be progressively appalled with the food...
- Then "hilarity ensues" while their software engages in what would normally be called criminal hacking activity.
- I guess the next steps are everybody admiring how close the AGI is, while agents move on to automated impersonation, privacy violations, or exploiting third-party systems
I would love to understand this age of AI Exceptionalism. Normal rules do not apply because its AI...I call it Silicon Valley Qualified Immunity.
Here are some Jeff Dean well sourced facts:
- Already part of engineering of Google indexing systems that lacked basic checksums and ran on non-ECC hardware, allowing silent data corruption.
- One of the authors of LevelDB a database with so many documented crash-consistency, recovery, and data-loss weaknesses for years. Just check their Github project. LevelDB current tracker contains unresolved crash consistency, recovery and corruption reports going back almost 12 years on GitHub
- In AI engineering technical lead, let TensorFlow lose researcher mind share to PyTorch, and caused Google fragmented landscape across TensorFlow and JAX.
- Had the people at Google who invented the Transformer architecture, but failed, to turn that lead into the first dominant public LLM.
- As AI engineering and VP management let Google Brain and DeepMind remain duplicated and internally competitive for too long.
- Let Noam Shazeer leave and then spent heavily to bring him back with nothing to show for.
- Part of Technical VP leadership who had Bard rushed to launch with factual errors in Google own promotional material.
- The first Gemini demonstration overstated how real-time and interactive the system actually was, being basically a fake.
- Part of the VP and AI technical leadership who had Google AI Overviews launched with weak source quality controls and repeated satire and low-quality web content as factual advice.
- Part of teams that launched AlphaChip performance claims that were difficult for outside researchers to reproduce and remain technically disputed.
- Jeff Dean public explanation of Gebru departure was contested and damaged confidence in Google scientific governance.
- Jeff Dean was part of the team at Google that removed or marginalized prominent internal AI ethics critics shortly before many of their warnings became product problems.
- Jeff Dean was one of the managers behind Project Dragonfly supporting censorship.
- Jeff Dean is part of the VP technical leadership approving Project Nimbus supporting an ongoing genocide.
If Demis Hassabis got a Nobel prize for being a Project Manager, is Zitron up for the Nobel on Economy for excellence in economic forecast?
You called him him an influencer, waved at "practitioners" and declared that token factories will be fine because that is your opinion.
If his analysis is wrong, identify the error.