ハル (@PatrickHalstead) header
ハル (@PatrickHalstead) profile picture

日本は人類の極相林。私は人間文化の社にささやかな供物を奉りに来た者にすぎません。3年前にシアトルから移住。三回目のデートで駆け落ちした愛妻は36年ものでキープ中。週50キロ走る25年もの。四国遍路結願、慈尊院から高野山まで町石道を五十回、般若心経を日課に。万年MVPの小さなソフト会社を細々と23年。

神園
Joined December 2010
Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)Photo by ハル (@PatrickHalstead)
Introducing The AI Scientist: The world’s first AI system for automating scientific research and open-ended discovery! sakana.ai/ai-scientist/ From ideation, writing code, running experiments and summarizing results, to writing entire papers and conducting peer-review, The AI Scientist opens a new era of AI-driven scientific research and accelerated discovery. Here are 4 example Machine Learning research papers generated by The AI Scientist. We published our report, The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery, and open-sourced our project! Paper: arxiv.org/abs/2408.06292 GitHub: github.com/SakanaAI/AI-Scien… Our system leverages LLMs to propose and implement new research directions. Here, we first apply The AI Scientist to conduct Machine Learning research. Crucially, our system is capable of executing the entire ML research lifecycle: from inventing research ideas and experiments, writing code, to executing experiments on GPUs and gathering results. It can also write an entire scientific paper, explaining, visualizing and contextualizing the results. Furthermore, while an LLM author writes entire research papers, another LLM reviewer critiques resulting manuscripts to provide feedback to improve the work, and also to select the most promising ideas to further develop in the next iteration cycle, leading to continual, open-ended discoveries, thus emulating the human scientific community. As a proof of concept, our system produced papers with novel contributions in ML research domains such language modeling, Diffusion and Grokking. We (@_chris_lu_, @RobertTLange, @hardmaru) proudly collaborated with the @UniOfOxford (@j_foerst, @FLAIR_Ox) and @UBC (@cong_ml, @jeffclune) on this exciting project.
282
1,453
727
6,001
3,433,130
#拡散希望 週プレNEWS @shupure_news の連載コラム第49話 「新型コロナウイルス学者」、ついに新型コロナウイルスに感染する …ついにこの日が来てしまいました😭 今回のコラムは私の感染闘病記😷 現在も流行拡大フェーズに入りつつあります。みなさん気をつけましょう🔥 wpb.shueisha.co.jp/news/tech…
17
505
31
861
138,722
When we started AWS, we looked at the undesirable situation of developers spending 70% of their time on the undifferentiated muck of managing infrastructure (vs only 30% on what differentiated their customer experiences) and aimed to flip that equation on its head. It’s made a pretty big difference in innovation and developer productivity. Roughly 18 years later, we’re trying to solve another 70-30-like misalignment for developers, where developers are telling us that they’re spending roughly 70% of their time on repetitive and tedious tasks and code. Today, we’re launching the general availability of Amazon Q, the world’s most capable GenAI-powered assistant for accelerating software development and leveraging companies’ internal data—and that aims to remove much of the muck of repetitive and tedious coding and data tasks for developers and employees at large. On the software development side, Q doesn’t just generate code, it also tests code, debugs coding conflicts, and transforms code from one form to another (today, developers can save months using Q to move from older versions of Java to newer, more secure and capable ones; in the near future, Q will help developers transform their .net code as well). Q Agents does multi-step planning and reasoning to allow developers to string together multiple requests and have Q implement them. On the internal data side, most companies have voluminous troves of internally relevant data that resides in wikis, intranet pages, Salesforce, storage repositories like Amazon S3, and a bevy of other data stores and SaaS apps that are hard to access. It makes answering straightforward questions about company policies, products, business results, code, people, and many other topics hard and frustrating. Q makes this much simpler. You can point Q at all of your enterprise data repositories, and it’ll search all this data, summarize logically, analyze trends, and engage in dialog with customers about it. We also introduced today a powerful new capability called Q Apps, which lets employees describe, in natural language, what apps they want to build on top of this internal data, and Q Apps will quickly generate that app. This is going to make it *so* much easier for internal teams to build useful apps from their own data. Customers are gravitating to Q, and we already see companies like Brightcove, British Telecom, Datadog, GitLab, GoDaddy, National Australia Bank, NCS, Netsmart, Slalom, Smartsheet, Sun Life, Tata Consultancy Services, Toyota, and Wiz using Q. And, we’ve only been in beta til today. Very excited about how Q will change what’s possible for our customers, and being a part of helping them innovate more quickly. aboutamazon.com/news/aws/ama…
81
432
110
2,917
663,513
ハル retweeted
My guess is we end up defining intelligence like we ended up defining arithmetic ability.. something that humans can do but machines can do better. AI becomes like forklifts, augmenting one human to do the work of many. Just as the stout working man of the 19th century gave way to the effete office dweller of the 20th The intellectual titan of the late 20th will give way to the emotive aesthete of the 21st Intention and motive will predominate skill and knowledge. Perseverance will continue to be of high import, especially given the number of seductive cognitive traps that the new new new economy will have to create. If you can hold an intent for a long period, direct other intelligences to carry it out, and compound over time, you will be supremely powerful.
14
12
127
8,981
If everyone wore a well-fitted respirator, the pandemic would be over in a few weeks. That's as controversial as saying a cholera epidemic can be ended by boiling/filtering all drinking water. No, there is no RCT for either of those claims—nor should there be. #COVIDisAirborne #CholeraisWaterborne
5
74
7
232
11,162
With respect, Dr Jha, it took us a while to discern long term effects of measles. Remember that for effective public health comms, we have to embrace uncertainty. Your messaging will further damage trust, esp w new data showing there is indeed Covid-induced immune dysregulation.
There's a lot of bad information out there about how covid damages the immune system It really doesn’t But you know what does? Measles And thanks to the antivaxers We are seeing measles outbreaks again Encourage people to get their kids vaccinated It’s really important
1
39
4
268
19,012
You sometimes lose hope in humanity, but then you meet people like those from Cleairwin. When I first met the founder, I asked him why he wanted to bring Flo Mask to Japan. It was simple, he said I just want to help people. I love it when our missions are aligned. Below is proof they're the real deal. Sending lots of love to Japan ❤️
【体育館避難所の盲点】 天井が高く暖房が効かない事です。 5m位あるので暖い空気は殆ど人の居ない上の方に溜まります。 クレアウィンBOXは最大1100㎥/hの風量の「サーキュレータ」搭載なのでこの上熱を下に下ろして来ます。 田鶴浜高避難所リーダーはこの事も喜んでおられました。 #強制対流
2
48
6
157
14,585
Oncology appt yesterday. There is still hope despite stage IV cancer and chemo ahead. #longcovid ravages lives. Kindly RT and contribute to my @PandemicPatient fund raise to help others. secure.givelively.org/donate…
9
74
3
185
17,187
Furthermore, JN.1 shows a robust resistance to monovalent XBB.1.5 vaccine sera🔥 These suggest that JN.1 is one of the most immune-evading variants to date. S:L455S can contributes to increased immune evasion, which partly explains the increased Re of JN.1. 7/7
7
101
23
185
34,694
Replying to @Gaijinotousan
The way of the Buddha is to know yourself; To know yourself is to forget yourself; To forget yourself is to be awakened by all things. — Eihei Dogen, quoted in Buddhism Without Beliefs by Stephen Batchelor
1
3
189
"As we get closer and closer to superintelligence, everybody involved gets more stressed and more anxious and we realized the stakes are higher and higher" We are further ahead than you think, this should be obvious by now.
Sam Altman explains what happened during the past few weeks at OpenAI
145
36
1,177
280,039
Everyone is using RAG, but most of the retrieved context is noisy! 🚨 Introducing FilCo: “Learning to Filter Context for Retrieval-Augmented Generation” TL;DR: Get rid of the irrelevant content using FilCo, and you'll get better outputs. Preprint: arxiv.org/abs/2311.08377
6
54
8
281
51,140
How do you perform quantum computation? Take a deep dive into quantum state concepts and probability calculations in the best environment to perform them: a Wolfram notebook. Check out this computational essay on #WolframCommunity to learn more: wolfr.am/1hWGkUTyT
6
31
4,493
It’s over. Wow. Wouldn’t have seen this coming 2 or 3 years ago. – at West Menlo Park, CA
BMW, Mini, Rolls-Royce, Toyota, and Lexus are switching EV plugs trib.al/sdMSar2
3
2
11
6,008
BREAKING🔔 A new study from G2P-Japan🇯🇵 is out! The preprint will be out so soon, but first we show our results on BA.2.86 - transmissibility, infectivity and immune resistance. This would be the 4th non-peer-reviewed🗒, but includes some new stuff. Please retweet/repost🔥 1/
28
635
109
1,288
674,721
Just to check we are all on the same page on neutralization and immune evasion - as the question is not if but why BA.2.86 is spreading, If it's not due to immune evasion, that's the less preferred scenario.
19
54
10
335
116,136
6/ Conclusions 📍Cardiovascular & neurological symptoms of #LongCOVID are deep & wide 📍People cont to get infected & develop this awful syndrome 📍Let’s stand w them & find medical answers via well-designed studies Here’s our new Nature Immunol paper: go.nature.com/3Pai6Oe
9
75
3
321
21,728
Replying to @paulg
Two things 1) Area under the curve matters and represents many who died preventable deaths in Sweden who didn't die in Canada, Japan or Australia in the same time-frame 2) Why are excess deaths still happening at all, 3.5 years out? That's the big question no one is asking.
9
17
1
86
8,636
BA.2.86 (Pirola) is in Switzerland wastewater. If we had good wastewater surveillance throughout the world, we would have a much better idea of how widespread this is. Unbelievable that funding for wastewater surveillance is being slashed in many countries.
BA.2.86 has been detected in wastewater samples from Switzerland (Laupen; canton Bern) taken on Aug. 5 and 6. Its frequency is rather uncertain, somewhere around 2%. The other 13 monitored locations across the country do not show a signal for BA.2.86-defining mutations yet. 1/2
56
320
31
1,274
225,342
What a week for public health 🤦🏻‍♂️ I don’t know who needs to hear this, but: surviving an infection doesn’t make you “stronger” for it. Surviving a #COVID19 infection leaves a mark. Like a boxer, every fight leaves it’s mark on your body. It can be an easy fight or a hard fight 1/
52
783
95
2,446
202,927