XERV (@waltcob) header
XERV (@waltcob) profile picture

software engineer | saxon

New York, NY
Joined February 2013
Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)Photo by XERV (@waltcob)
📌
"You got lucky" The luck:
1
4
4,826
hi Thariq (@trq212), i got a beautiful tool for this.
I asked Claude how I could save space on my computer
33
15
4
1,168
72,835
XERV retweeted
Easily one of the funniest deaths of all time
Anyways, here is Clark watching his father die from a tornado💀
169
4,227
73
95,448
1,661,176
Marx failed to consider the devaluation of labour by machinery
Realizing that leftists hate generative AI so much because it obliterates the labor theory of value.
38
622
17
11,854
170,862
the ultimate monorepo
63
58
9
1,727
94,926
XERV retweeted
Printf in C is so BAD I had a Crash Out
63
65
16
1,729
64,497
Oriel College, Oxford University: Racists Welcome Here. Words: @kehinde_andrews
366
908
48
2,598
250,934
BREAKING: New analysis by mathematician Jenny Lorraine Nielsen claims the AI-assisted proof of the Riemann Hypothesis is invalid, despite its successful verification in Lean.
213
199
77
3,348
363,498
Anthropic solved the biggest unsolved problem in math for 167 years and it’s just a butthole
Riemann Hypothesis
27
53
3
1,059
34,486
On Transgender Day of Remembrance, we mourn every person stolen by anti-trans violence. Their lives mattered. We honor their memory by fighting for a City where every trans New Yorker can live in safety and freedom. That means protecting access to gender-affirming care and standing up to those who would make our neighbors less safe. Our commitment to that fight will not waver.
6,329
31,173
974
183,916
2,530,866
Open-source desktop application that records, transcribes, and analyzes meetings using real-time AI. github.com/video-db/call.md
9
118
5
1,090
48,880
XERV retweeted
Believe me Interview process has become harder than the actual job
32
6
1
160
8,756
YOU CAN PUT YOUR LLM IN A LAMBDA FUNCTION! TELL EVERYONE! LAMBDA FUNCTIONS CAN NOW RUN ON GPUS!
144
232
100
4,400
348,137
it’s more easy than ever to have an accidental 1M bill from AWS technology is beautiful
YOU CAN PUT YOUR LLM IN A LAMBDA FUNCTION! TELL EVERYONE! LAMBDA FUNCTIONS CAN NOW RUN ON GPUS!
22
175
2
4,012
51,875
Sarah Silverman: 'I hope the Jews did kill Christ; I’d do it again. I’d f*cking do it again in a second'
93
257
80
717
271,768
this needs to be a federal law immediately.
🚨 Governor Josh Shapiro signs bill forcing advertisers to disclose when AI is being used.
36
552
3
10,571
85,726
New NVIDIA paper finds that once an LLM has reasoned through a problem, the solution can run as CPU code for rephrased and re-numbered versions of it. that answers similar questions 16x to 291x cheaper. LLMs re-solve every problem from scratch, even when only wording or numbers change. Agents sending the same instructions with new data pay for that repeated reasoning each time. Their method has Claude Opus 4.5 turn a reasoning trace into small Python functions, with changeable inputs as parameters. At run time they execute on a CPU, calling a fine-tuned 1B model only for small text jobs like extracting values. The cache matched or beat the source model's accuracy on 7 of 8 math and science benchmarks, scoring 93.3% versus 83.3% on AIME '25. It handled at least 90% of rephrased or re-numbered variants. If your agent runs the same prompt on new inputs all day, cache the reasoning as code and use the LLM only on misses.
69
166
35
1,415
54,964
XERV retweeted
helping others for just 2-4 hours a week was linked to 15-20% slower brain aging.
6
18
1
221
5,029
This video 🤪
143
1,826
77
9,640
362,454
XERV retweeted
China open-sourced a peanut-sized OCR that parses entire 100-page PDFs in one shot.. It's called Unlimited-OCR. Only 3B params. Runs locally. Every other OCR tool chops your doc into pages and loses the thread. this one reads the whole thing in a single pass. → One-shot "long-horizon" parsing (32K context window) → Multilingual, out of the box → 93% on OmniDocBench v1.5 (+6 over the DeepSeek-OCR baseline) → Edit distance stays under 0.11 past 40 pages → Runs 100% locally on your own hardware → Works with Transformers, vLLM, SGLang, and Docker Traditional cloud OCR (Textract, Google Vision, Azure Document Intelligence) often costs about $1.50–$15 per 1,000 pages. This runs on your machine. For free. Forever. Baidu built it to push DeepSeek-OCR one step further. It now has millions of downloads on Hugging Face, and most people still have no idea it exists. 100% open source (MIT). github.com/baidu/Unlimited-O…
22
166
12
1,774
126,636
Unspoken expectations are the death of relationships.
71
273
17
2,056
92,920