Latest AI/ML News
7510 matching items
미 태양광 제조사들, 한국산 셀 우회조사 청원…한화큐셀 포함
한국산 태양전지가 미국 태양광 반덤핑·상계관세 논쟁의 새 쟁점으로 떠올랐다. 조사가 실제로 시작되면 한국산 셀을 쓰는 미국 모듈 공장과 태양광 발전소 조달 비용에도 영향을 줄 수 있다는 분석이다.로이터는 22일(현지시간) 캐네디언 솔라(Canadian Solar), SEG 솔라(SEG Solar), 헬리엔(Heliene)이 미 상무부에 한국산 태양전지 수입에 대한 우회조사를 요청하는 청원을 제출했다고 보도했다.이들은 한화큐셀 등 한국 셀 생산업체가 중국산 태양광 제품에 적용되는 기존 반덤핑·상계관세 명령을 우회하고 있는지 조사해야
Recipe Rescaler
I keep my recipies on my website , and like most of my website it grew over time instead of being designed. A couple years ago I added some progressive enhancement that puts checkboxes on the ingredients , and today I added a rescaler: Here's tripling it: This is another project, like adding transposition to my solstice songbook , where I wouldn't have put in the time if I couldn't delegate to an LLM. It went very quickly, and the code seems reasonable. Implementation notes: As you go up and down it converts teaspoons to tablespoons to cups. Yes, I still cook volumetrically. It handles numeric ranges, like "3-4 cups". It handles fractions: half of 1 1/4 C is 5/8 C. It doesn't handle everything. I wanted something simple and reviewable that handles most cases, instead of trying to make something exaustive (that would then have weird bugs). This means with complex items like "2 eggs (or 2T flax and 5T water)" only the "2 eggs" is scaled. To make these failures graceful, all scaled values are bolded, so unscaled values stick out visually. Comment via: facebook , lesswrong , mastodon , bluesky Discuss
정부, 송전용 BESS에 그리드포밍 성능 요건 적용
재생에너지 확대에 맞춰 송전용 에너지저장장치(BESS)의 역할이 전력 저장에서 계통 안정화로 넓어진다. 태양광과 풍력 등 인버터 기반 발전설비가 늘어나면서 전력망의 전압과 주파수를 안정적으로 유지하는 기능이 중요해졌기 때문이다.기후에너지환경부는 최근 그리드포밍 성능 요건을 마련하고, 2027년 12월부터 상업운전에 들어가는 중앙계약시장 장주기 BESS 설비부터 적용한다고 밝혔다.그리드포밍은 인버터 기반 설비가 전력망의 전압과 주파수를 스스로 만들고 유지하도록 제어하는 기술이다.기존 그리드팔로잉 방식은 이미 형성된 전력망 신호를 따라
Israelis, Palestinians torn over sacred shrine in city of Hebron
Israelis, Palestinians torn over sacred shrine in city of Hebron
Celebrating 1 Million Members and Introducing the Community Leaderboard
Wow, I haven’t even been here for 2-3 months yet, but even I’m ranked 24th on the list. I like that. Edit: I just realized the ranking I was looking at was already the monthly ranking When looking at the yearly ranking, I’m 104th. All-time, I’m 422nd. Still not bad
Windows app briefly becomes unresponsive
(post deleted by author)
Australia investigating five social media giants for not enforcing ban on kids
PLUS: Qualcomm tweaks datacenter chips for China; Japan's Air Force stretches into space; WiseTechs' woes mutiply; and more!
Switch for gpt image gen in long context prompts lead to "Connection interrupted. Waiting for the complete answer"
Since the last addition of GPT 5.6 for testing, the image gen system has been having constant trouble rendering images in the chat. In my case, about 30% of the generations are never rendered in the chat, leading to the message “Connection interrupted. Waiting for the complete answer” or “Image generation failed, try again”. Thanks for your valuable work, OpenAI team. I am bringing this to your attention because it has been happening for about four days in a row, with no interruption.
Iran warns ships not to bypass its chosen Hormuz route
Israel and Lebanon hail Washington deal, rejected by Hezbollah
Codex Analytics shows GPT-5.4 for most turns, but my CLI is configured to GPT-5.5 — why?
It is codex-auto-review. x.com Tibo @thsottiaux @solipsnitsyn Auto review is using GPT-5.4. We do not route. Do not spread misinformation, thank you. 10:31 PM - 28 Jun 2026 310 1
Paraguay eye 'life-changing' Germany World Cup upset
Paraguay eye 'life-changing' Germany World Cup upset
I built a whole-home ad blocker with a $7 ESP32-S3 board - and it took just minutes
Raspberry Pi boards have gotten expensive, so I found a cheap alternative in a tiny ESP32-S3 board.
FCEDA to Support Korean Tech Firms Entering U.S. Market
The Fairfax County Economic Development Authority (FCEDA) is a public, non-profit agency of the Fairfax County government that operates a wide range of support programs for local businesses.Fairfax County, Virginia, in particular, is a strategic location often referred to as a gateway to Washington,
RANDOM Codex weekly quota reset still happening. WHY?!?
It is not random. They are working on a sunday to keep things running. x.com Tibo @thsottiaux Codex team is in a warroom on a Sunday combing through logs and checking whether there is anything that could lead to increased usage drains for some users. Taking it very seriously and won't rest until we get to the bottom of it. x.com/nicdunz/status… nic @nicdunz i think openai has gotten a bit too loose and funny acting. i think its time to get a little serious 10:18 PM - 28 Jun 2026 2.3K 61 x.com Tibo @thsottiaux If you happened to have used your previous reset in the few hours before and didn't go through your usage, do not worry, you will get more manual resets after we conclude the investigation. 12:01 AM - 29 Jun 2026 290 7
A reading list for generalists
I, along with many others in AI safety, believe there is a shortage of generalists in the community and that there exist many projects and efforts that by default will not happen unless they are owned by a strong generalist [1] [2] [3] . As someone who is a reasonably good generalist, I decided to assemble a reading list of the essays and blog posts that have personally helped me the most. I would love others to comment with pieces they think should be on this list. The crux of this reading list is the idea that if you’re working hard as a generalist on a project you care a lot about, then by rigorously applying the lessons from these documents you will improve more quickly than you otherwise would. By the numbers: I’ve attached 18 documents to start this reading list. The authors cited more than once are Paul Graham (5), Ben Kuhn (4), Ethan Perez (2), and Greg Brockman (2). Sam Altman and Eliezer Yudkowsky also have their fingerprints over a lot of the content. The items are 15 blog posts, 1 blog comment, 1 interview transcript in blog post form, and 1 book. Dispositional What characteristics should you try to adopt? Paul Graham: "What We Look for in Founders" ( link ), "Relentlessly Resourceful" ( link ) Eliezer Yudkowsky: "Shut Up and Do the Impossible!" ( link ) Ben Kuhn: "Be impatient" ( link ) Cate Hall: "How to be more agentic" ( link ) Strategy How do you make good decisions with the information you have, and how can you get the additional information you need? Anna…
Third-parties should focus on scrutinising system cards
By default, I expect system cards will get worse, which would be bad. Some mechanisms could improve system cards, but I expect they will be outweighed. In any case, I think third-parties should focus on scrutinising system cards — this seems like a great activity for outsiders in the current strategic landscape. I'll sketch what that could look like, and offer some recommendations. It would be bad if system cards degraded. If labs felt pressure to evaluate the risks accurately, they'd be better incentivised to reduce them. If the risks were high enough, and a lab communicated that, then this might prompt drastic government action. It's very plausible that, if labs build misaligned AIs that take over, then most of the employees had a genuine but incorrect belief that the AIs wouldn't take over, based on evidence that was actually flimsy and misleading. Outsiders scrutinising system cards seems like a great mechanism for improving the epistemics within the labs. It's good for the outside community to know which risks are most pressing, so they know which activities to prioritise . By default, I expect system cards to get worse, because… The system will get genuinely more complicated. There will be more models, trained with more techniques, interacting in more ways — plus dozens of ad-hoc patches for the problems that arise. Already no one person holds the whole system in their head, and the fraction that fits in one head will keep shrinking. Any overall safety judgment will co…
Critical rescue window closing in Venezuela as quake death toll nears 1,500
Twin earthquakes in Venezuela destroy buildings, sow panic
I grew up in China, but I'm raising my daughter in LA. A trip back to visit family made me rethink my parenting style.
A visit home to China made Grace Cong Sui realize that the parenting habits she'd adopted in the US differed from those she grew up with.
I'm a stay-at-home mom of 2. My 4 side gigs help me pay off debt and stay sane.
Lana Ng left her marketing job in 2023 to take care of her children. Now, her four side gigs help her not lose herself in motherhood.
AI Weekly Issue #509: AI Productivity: it works best for the people losing their jobs
Three years into the productivity promise, there's finally enough hard evidence to answer the question plainly: does working with AI actually make you more productive? Yes — spectacularly, for some people on some tasks. And no, or worse, for others. The gains are real. They're just not flowing where the marketing said they would. This issue maps who wins, who pays, and why the line between them isn't where you think.
Documentation as Code: A Practical Guide for Dev Teams
Documentation as code keeps docs in the repo and runs them through the same review and CI as code, so they stay in sync. The workflow, the honest trade-offs, the tools, and how to keep docs true to the code.
SpaceX launches 24 Starlink satellites
SpaceX launches 24 Starlink satellites
Powerful qualities in type B physicalism
Type B physicalism is a view in philosophy of mind, and "powerful qualities" (PQ) theory is a view in the metaphysics of properties. They share an important similarity: they both combine conceptual dualism with ontic monism. Type B physicalists hold that phenomenal concepts (e.g. concepts of reddish experience) are not analytically definable in terms of physical concepts, yet can co-refer ("pain is C-fiber firing" being a simplified example), and that every property is a physical property. PQ theorists claim that there are both qualitative concepts (referring to qualities such as the redness of a red object) and dispositional concepts (referring to tendencies such as the disposition to reflect ~700nm light), and that every property is both really qualitative and really dispositional; hence, qualitative concepts co-refer with dispositional concepts. I suggest that type B physicalism combines well with powerful qualities theory. First, generalizing phenomenal concepts to qualitative concepts makes phenomenal concept theory less ad hoc . Second, considering "phenomenally conscious" as a secondary qualitative concept, and considering un-conscious qualities to exist, yields a more comprehensible theory. Third, philosophical debates about the two views are similar, and type B physicalists find themselves in a similar dialectical place to PQ theorists. Background I take type B physicalism to be the conjunction of three theses: Physicalism : There are no non-physical mental properti…
A family traded Connecticut for retirement in a car-free Costa Rica beach town
"We fill every day. There's really not a lot of downtime, yet we never feel rushed or pushed or stressed," Andrew Rappaport told BI about life in Costa Rica.
France's deadliest skydiving plane disaster kills 11 moments after takeoff
11 killed in skydiving aircraft crash in north-eastern France
2026年に採用が最も難しいIT職種11選——何が変わったか
専門家の採用はむしろ容易になった。SOCアナリスト、ML研究者、クラウドアーキテクト——こうした職種は数週間で採用が決まる。一方、採用に6〜9カ月かかるのはハイブリッドな職種だ。AIに精通しながらコードにも深く入り込め、ビジネスも理解できるエンジニア。「3つのスキル、1人の人物、少ない候補者プール」とBest BuyのCDTOであるNeal Sample氏は言う。「これらのハイブリッド人材がITの未来だ——現時点でこのようなハイブリッド人材は見つけるのがとても難しい」。 AIがサイバーセキュリティを抜いて採用最難関スキルになって2年が経過する。2026年のState of the CIO調査では、AI/機械学習とサイバーセキュリティが同率1位となり、データサイエンス・分析が僅差で続く、という結果になった。ランキングは似ているが、人材難の性質は変わった。LLMエンジニアやプロンプトスペシャリストへの需要は、AIをスケールで実用化し、リスクを管理し、盲目的に信頼せずに使いこなせる人材への需要に変わった。 リスク管理は初めてトップ5入りし、ビジネス・IT自動化は上位を維持している。一方で数年前まで注目されていた職種への需要は緩んでいる。その1つがクラウドアーキテクチャだが、今年は順位を落とし、アプリケーション開発もリストから外れた。「採用が最も難しいのは、AIとの組み合わせが求められる職種すべてだ」とValcom TechnologiesのITアドバイザー、Niel Nickolaisen氏は言う。 採用困難なIT職種：2026年 vs. 2024年 スキル 2026 年 2024 年 変化 AI/機械学習 1位（同率） 1位 横ばい サイバーセキュリティ 1位（同率） 2位 上昇 データサイエンス・分析 3位 3位 横ばい ビジネス・IT自動化 4位 4位（同率） 横ばい リスク管理 5位 8位（同率） 上昇 ソフトウェアエンジニアリング 6位（同率） 6位（同率） 横ばい DevOps/DevSecOps 6位（同率） 11位（同率） 上昇 エンタープライズアーキテクチャ 8位（同率） 10位（同率） 上昇 クラウドサービス・統合 8位（同率） 12位（同率） 上昇 クラウドアーキテクチャ 8位（同率） 6位（同率） 低下 デザイン思考・UX 8位（同率） 15位（同率） 上昇 出典：Foundry/CIO.com State of the CIO Survey、2024年および2026年 AI採用の成熟 「プロンプトエンジニアリングは単独の職種としては短命だった。現在はベースラインのスキルになっている」とSample氏は言う。多くの組織が求めるのは別のものだ——エージェントを立ち上げ、テストフレームワークを構築し、コスト・レイテンシ・品質のトレードオフを管理し、AIをスケールで展開できるAIプロダクトエンジニアだ。3年前は存在しなかったガバナンスやレッドチームの役割も生まれている。「モデルを作る人からモデルを使いこなす人へ、重力の中心が移った」とSample氏は言う。 このように、求められるスキルは、プロンプトエンジニアリングよりもエージェンティックAIの活用へとシフトしている。「ワークフロー、プロセスの簡略化を理解し、エージェントプラットフォームで業務を自動化できる人材が必要だ」とNickolaisen氏は言う。課題は、AIが急速に進化しているため、ある企業での経験が別の企業で通用しない可能性があること、また6カ月前に学んだことがすでに時代遅…
ITを揺るがす「エンタープライズ価格大転換」の波
SaaSおよびAIソフトウェア市場は価格の大転換期に入っている。従来型のSaaSと競合するAI製品の登場により、SaaSベンダーはシートベース（ユーザー数基準）の価格設定の見直しに迫られている。見直しの結果、従来型SaaSの価格はアウトカムベース（成果連動型）課金へと移行しつつある、これはCIOとCFOにとっては概ねプラスの動きだ。一方、一部のAIツールやSaaSパッケージは消費量ベースの課金へと向かっており、利用状況を注意深く管理しなければ、請求額が想定外に膨らむリスクがある。 CRMデータプラットフォームプロバイダーTwilioの戦略・オペレーション担当ディレクター、Sidharth Ramsinghaney氏は「最も差し迫った問題は、ほとんどの組織がこれまで経験したことのない予算の変動だ。この転換は、ベンダーから買い手に予測リスクを移転するものだ」と言う。Gartnerによれば、2030年までにエンタープライズSaaSの支出の少なくとも40%が利用量、エージェント、アウトカムベースの課金に移行し、シートベースのベンダー収益シェアは21%から15%に低下すると予測されている。 AIエージェントは価格モデルを揺るがす可能性を持っている。「AIエージェントが人間の業務を代替するようになれば、この予測は現実になる。以前は10人が必要だった仕事を1つのエージェントがこなせるなら、ユーザー数に基づくシートベースの課金は意味をなさなくなる」とRamsinghaney氏は言う。ただし、エージェントの導入が停滞すれば、シートベースの価格設定が長く続く可能性もある。不確実性が続けば、ハイブリッド型の価格設定がデフォルトになる可能性もある。 アウトカムベース課金の難しさ 一部の顧客がアウトカムベースの価格設定を求め、一部のベンダーが試験的に導入している。だが、実現は容易ではない。「アウトカムの帰属を明確に測定することが本質的に難しい」とRamsinghaney氏、「そのため、純粋なアウトカムベースの価格設定は多くのカテゴリーでまだ理想論にとどまる」と続ける。ソフトウェアコンサルティング会社Software Improvement GroupのCTO、Jasper Geurts氏も同意する。多くのAIベンダーはアウトカムではなく、シートベースと利用量の組み合わせに動いている。「エージェントはテストに合格しながら、誰も後から手を加えられない質の低いコードを生成することがある。機能すればいいという成果のみに課金するなら、将来の負債となるコードに対してお金を払っているのと同じだ」とGeurts氏は言う。 トークン経済の不透明さ AIベンダーのコスト構造は従来型SaaSとは異なる。フロンティアAIモデルのコストは下がるどころか上がっており、AnthropicのFable 5モデルはOpus 4.8の2倍のトークン単価だとGeurts氏は指摘する。「CIOたちはトークン経済が不透明だと言う。エージェントがどれだけトークンを消費するか予測できなければ、請求額の見通しが立てられない。CIOはトークンを10年前のクラウド支出と同様に扱うべきだ——つまり計測し、管理し、成果と結びつける必要がある」。 「シートベースの価格設定はヘッドカウントで請求額を抑えられた。利用量ベースの価格設定は消費量に応じてスケールする。AIがアシスタントから自律的なコーディングに移行するにつれ、コードの量はレビューしきれないほど増える。誰もレビューしきれないコードが積み上がれば、それ自体が将来の負担にな…
RTA expands strategic partnerships with leading Chinese companies
RTA expands strategic partnerships with leading Chinese companies
Putin acknowledges ‘difficult period’ as Ukraine ramps up attacks
Kyiv struck a Russian defense plant and set fire to a major oil refinery as part of its efforts to weaken Moscow’s wartime economy.
South Africa braces for wave of anti-migrant violence
Thousands of migrants are fleeing South Africa as the country braces for violence after an anti-immigrant group set a Tuesday deadline for all undocumented migrants to leave.
AI spending boom could end in global bust, report says
The AI spending spree risks an extended “investment bust” that could imperil the global economy, a report warned Sunday.
Care externalities
A big source of externalities I’m not sure I’ve heard mentioned as such: people caring about one another. For a random person, I’d guess the unaccounted for effects on other people from their making different private decisions about their lives while being cared about by numerous family and friends might be at least comparable to those from marginally contributing to air pollution or city density or traffic congestion. Discuss
Heat wave grips Europe, reigniting debate over AC
Austria, the Czech Republic, Germany, and Poland sweat through record temperatures on Sunday, as a European heat wave reignited a political debate over the continent’s preparedness for hotter summers.
Plane crash in Beijing reveals airspace security gaps
A small plane crash in China’s capital exposed major gaps in the country’s airspace controls, analysts said, as the government tried to limit public discussion of the incident.
US and Iran test truce with fresh strikes
Tehran attacked American military sites in Kuwait and Bahrain after the US accused Iran of attacking a cargo ship in the Strait of Hormuz and targeted Iranian infrastructure in response.
Quoting Jon Udell
Human Agent in the loop I dislike the phrase “human in the loop” because it cedes authority to the machines. Let’s flip the narrative. It’s our loop, we work the same way we always have, now we recruit agents to join the team. An agent-assisted process need not be a black box that takes in prompts and emits features. [...] Let’s do agentic software development like that. Not as a loop we’ve been excluded from, instead as one we invite agents into. — Jon Udell , “Doctor, it hurts when agents create unreviewable PRs.” “Don’t do that.” Tags: jon-udell , coding-agents , generative-ai , agentic-engineering , ai , llms
Arab League condemns continued Iranian attacks on Bahrain, Kuwait
Arab League condemns continued Iranian attacks on Bahrain, Kuwait
China’s Z.ai claims it can match Mythos on cybersecurity
China's Zhipu AI (Z.ai) released its open-weight GLM-5.2, and some researchers have claimed that it matches Mythos in certain bug-finding and cybersecurity scenarios. While GLM lags behind models from Anthropic and OpenAI in other, more general tasks, it seems that China has dramatically reduced the gap in the capabilities between its models and those of […]
California law targeting loud streaming ads takes effect on July 1
Streaming ads might be getting a lot quieter.
OpenAI support no response 3 weeks after june false ban wave
I don’t want to have to keep creating new topics about this @OpenAI_Support please let me know of any updates, i still have not received any response in almost a month since it was escalated to a “Specialized Team” and I haven’t gotten any updates here. Is anyone still looking at my case or what is happening?
Stars, Stripes, VIPs and Protests: Brussels hosts US Independence bash
Among those attending are Belgian Prime Minister Bart de Wever, NATO Secretary General Mark Rutte, and European Parliament President Roberta Metsola.
Windows codex-app is completely messed up
the latest app version: Frequently crashes CLI chats / instance are not synced with UI, CLI chat do not appear in the app unable to transcribe furthermore, for remote control, the one QR code per account is useless feature if you have multiple accounts due to rate limits. so when one account is out of limit and you switch to antoher account you have to reset up your remote control in the app. account 1 sign in - get codex coding, sign in to remote control using QR code, runs out of limit, switch to account two, now need to resign in to use remote control with new QR code.
Joy as man pulled from rubble three days after Venezuela quake
Spanish and international rescue teams pulled a survivor from the rubble four days after deadly earthquakes struck Venezuela. Search operations continue as crews race to find more survivors, with the death toll rising to more than 1,400.
The battle for the worst job in Britain heats up
Andy Burnham’s choice of chancellor will be the clearest signal of the direction he is planning on taking the country as a whole.
Suno launches Spark incubator program to feed independent artists to its AI machine
Suno has ambitions to be more than just a toy to churn out AI slop, it also wants to be a streaming destination and to break new artists. Spark is their new incubator program for independent artists that provides grants, mentorship, and marketing support. To apply, artists need to be an unsigned singer, songwriter, or […]
The problem of how to gain support
All right this is getting ridiculous . 3 weeks I am getting messages that support is looking into this but solution is still not provided, can someone write to me what is the issue and can we get this resolved finally??