Panel
Iniciar SesiónRegistrarse

Noticias Financieras

3 ASX 200 shares I would buy in February
fool_auhace 63d

3 ASX 200 shares I would buy in February

Here's 3 quality ASX 200 shares to watch this month as market conditions shift.The post 3 ASX 200 shares I would buy in February appeared first on The Motley Fool Australia.

#COMMODITIES
The Fed, Electricity, And Affordability
zerohedgehace 63d

The Fed, Electricity, And Affordability

The Fed, Electricity, And Affordability By Peter Tchir of Academy SecuritiesThe Fed, Electricity, And AffordabilityThree distinct topics, but at the same time, they are almost (kind of) the same topic, or at least interconnected.A Warsh FedWe discussed Warsh on Friday (along with gold, Bitcoin, and the flippant use of “debasement”). Dismissing the “Warsh is a Hawk” narrative:We are looking for 3 rate cuts by September. Not 300 bps, just 75 bps. That isn’t a heavy lift.It is easier to be hawkish when you are not the person who risks sinking the economy into a recession. We’ve argued for ages that whoever becomes the Fed Chair immediately shifts two notches more dovish. Yes, inflation is painful, but by definition, it typically takes time, and is more of a “slow bleed” that is often masked in the early stages. Recession hits pretty hard, pretty quickly. You really think the Fed Chair errs towards fighting inflation rather than keeping the economy running smoothly? (They got it wrong in the other direction once). My working assumption is that it is more difficult to be the “inflation hawk” when you are going to get the primary blame for tanking the economy. When the President (and Bessent and Miran - more on him in a moment) all want you to cut. When your wife’s father is a pretty large donor to the admin. Imagine that Thanksgiving dinner conversation. “You didn’t cut, we lost the midterms because of that, please pass the gravy.”Miran spoke on Friday, and I swear he has been reading the T-Report as he argued about the reality of the neutral rate being lower than the current Fed believes (in aggregate). There are valid arguments for cutting. More importantly, he argued, as we have for years, that housing in CPI is lagged. It tells us things from 6 months to a year ago – not today.When we all know (mathematically) that the data in CPI is wrong, why do we base decisions on it? We missed “transitory” because of this and we are at risk of missing a shift in inflation again.“Hey Joe, why are you going outside without an umbrella?”“Because my app says it’s not raining.”“But you can see the rain, and you can hear the rain pounding on the roof!”“Yeah, I’m going with the app, it’s probably right.”As stupid as that conversation sounds, that is where I feel we are on some of this data. We did hit on both OER vs Zillow and Truflation vs CPI in last weekend’s T-report.Reluctant to use the balance sheet. This one is tricky for me as I do believe QE tends to spur inflation. So, it is easy to see Warsh not wanting to use QE to control the yield curve. Having said that, Operation Twist is not viewed as balance sheet expansion by the Fed. Us mere mortals view Operation Twist as a form of QE because they take a lot of duration out of the market (selling bills to buy bonds). So that part of my view has not been removed. Even if he cuts, and the market becomes convinced that it was warranted (which I think it is), he might not have to do much to control the long-end of the curve. Will he do QE/Yield Curve Control? I’m less certain that is the ultimate endgame with him, but I still find it difficult to believe that we won’t take extraordinary measures to lower mortgage rates (more on that later), which are linked to the 10-year.Coordination and Cooperation are coming. Another theme from last week (and prior reports) is that we should expect more coordination between the admin, the Treasury, and the Fed. Warsh’s choice fits that narrative well.I think the market will come to terms with this, but I also think the “debasement” trade was so overdone, and there is more unwinding on that.ElectricitySame chart as last week.The cost of electricity and the inability to produce enough electrons and get those electrons to where they need to be has evolved into one of the most important discussions in this country, and it is rapidly becoming the primary topic in other countries (ZH: as we first said long before everyone starting piggybacking last August)In one year, this will be the most popular chart on this site pic.twitter.com/h93gWXMoNL— zerohedge (@zerohedge) August 11, 2025Everywhere you look there are bottlenecks:Getting permission to build electricity generation facilities.Getting the equipment and parts needed to build the facilities.Getting the fuel source to wherever you are building the facilities (uranium and solar have some advantages here).Getting the electricity to where it is being used. The grid leaves a lot to be desired. Positioning data centers that can tolerate higher latency, closer to energy sources, might be an alternative.I cannot begin to describe how “electricity” sucks the air out of a room right now. Instant attention from the audience. Questions, concerns, thoughts, ideas, etc. I’d be shocked that if you mention electricity bills to any 10 people, you will not find at least one person instantly engaged! The concern is bipartisan. The solution is ProSec.AffordabilityToday we will focus on “shelter” affordability. And maybe just a little on autos and healthcare.While we won’t go into detail here, I think we need to address the subject of the working poor.The Working PoorI am sick of the “K”- shaped economy. First, I think at best it is k-shaped (the upper leg is much smaller than the leg heading down), but it probably is more of an i-shaped economy (little I).This “letter shaped” discussion hides the ugly truth of the “working poor.” People with jobs who cannot make ends meet.This isn’t just people at the poverty level (the calculation is bizarre). These are people with “normal” jobs who have/had a “normal” life with that job. Or at least they did 5 to 10 years ago. Now they are struggling to keep up a lifestyle that didn’t seem to be a reach/stretch just a few years ago.I had a really interesting conversation with one of our clients and it really triggered that sort of a “eureka” moment.He discussed trying to prepare his business for a “working poor” recession (not quite his words). This was in contrast to an unemployment-led recession.We’ve had recessions caused by (or at least coinciding with) job losses. People lose their jobs and the economy heads into a recession.Are we at risk of having a recession because people just cannot keep up their lifestyles even while keeping their jobs and getting raises?I don’t have a strong view (honestly it just hit me this week during that conversation), but it meshes with a lot of concerns that I’ve had about consumption and the economy.We will certainly think about this more, but it immediately backs into why affordability is such a major issue. And it is less about the rate of inflation (intellectual fluff) than the high cost of living (inflation never captured how expensive our lives have gotten).Housing AffordabilityLet’s hit a few things here, but start with one premise.Lowering the average price of homes is NOT good. I lived the “Big Short” but hated the book and never got to the movie, because it made it seem like no one had a clue a bubble was forming. Lots of people saw the bubble, they just didn’t time it well enough to have capital left to risk when it all came tumbling down. If you remember the premise was along the lines of “the average price of homes in America has never gone down.” They did and we got the GFC. Crashing home prices isn’t going to help the economy or country. For so many Americans, their home is their largest store of value and I suspect not much has changed in less than 20 years, so I’d try to avoid driving prices down.The one we’ve already talked about:Electricity. Subsidies? Forcing hyperscalers to directly fund not just their energy needs, but also their communities? Who knows what is enforceable or plausible, but with the $$$ around hyperscalers they may make a “convenient” target for politicians looking for votes. The “pain” will be moderate, at least in the overall $$$ context, because we need the data center and AI growth to continue, but something around this could have widespread appeal in an election year.In the meantime, hopefully we can build our way out of this more rapidly than many (including me) think we can.One area we’ve touched on a bit in the past:Mortgage interest. The lower the interest payments are, the more “affordable” the house is.Reduce spreads on mortgages. Have agencies buy even more? Warsh would seem reluctant to do that with the Fed balance sheet, but this could happen outside his purview.Lower Fed Funds. This could provide some immediate relief for those willing to take on floating rate mortgages.Lower 10-year yields. The “Holy Grail” as it benefits mortgages while letting people get the comfort of a fixed rate rather than dealing with floating rate risks.50-year mortgages? I think it is a “suboptimal” idea. You don’t lower the mortgage payment that much while creating all sorts of new risks for the borrower and the lender.Portable mortgages? Some chatter about this, but that seems to add to inequality. Those with existing mortgages have an advantage in the market. I think this is a zero-sum game and not worth doing as it creates a lot of potential issues, while I struggle to see how it helps “create” housing. It might let some people move for job purposes, who feel stuck, but again, that is just shifting inventory around, not creating new inventory or reducing payments for someone else.Job growth in cheaper locations. Not every city or area in the country has the same cost of buying a house (or living there). Some are clearly tied to the types of jobs that a community can support, but there is room, I believe, to see that “reindustrialization” (or ProSecTM as I prefer) can create new jobs in areas where the cost of housing and living can be more affordable, mitigating the risk of getting stuck as “working poor.”For now, let’s treat this more “sensitive” subject as a corollary to jobs in cheaper locations. Venezuela and Mexican Cartels. There is ample reason to believe that Venezuela will be safer for the average citizen and that “normal” jobs (not drug-related jobs) will be created as investment in oil production (and rare earths/critical minerals) grows. That may cause some Venezuelan immigrants in the U.S. to return home. We haven’t yet seen any aggressive action against the Mexican cartels, but that is certainly on my bingo card ahead of the midterms. As many flee Mexico not just for jobs in America but also to avoid the horrible choice of “silver or lead” (join the cartel or get shot), we could see many return to Mexico if a better environment is created. U.S. companies would need clarity on tariffs, but they could invest in plants there too (again, in my vision of ProSecTM working with close neighbors and allies will play a role). For full disclosure, for “risk management” purposes, I’m starting the process of switching from a green card to citizenship. In any case, this could free up some housing availability in the U.S.Who’d have thought that moving to home insurance would be a “comfortable” step. Housing insurance increased 5% from 2014 until 2022. It is up 13% in 3 years! There are lots of reasons for this. The cost of repairs has increased. The time to do a repair has increased, which not only increases the direct cost, but it now also costs more for families that need to rent somewhere during repairs. These are market forces at work. Could the President “cap” insurance premium increases? This isn’t like Medicaid payments where the government is the payor, but on the other hand, could he cap credit card rates at 10%? I don’t think we should interfere with market forces, but I’m not the President, I’m not trying to win the midterms, and I wouldn’t cap credit card interest at 10%. As an investor, I’d keep an eye on this. As a lobbyist, I’d make sure the reasons for the increase are well understood and deemed fair. By the way, the auto insurance chart wasn’t as stable, but it has also grown rapidly.Things associated with the cost of owning a home (the mortgage, the insurance premium, the utility bills) will all likely be focused on by the admin in their effort to drive “housing affordability” lower.With auto ownership (including leasing) closely associated (at least in my mind) with home ownership, that is another area that could be identified by the admin for some special scrutiny in their efforts to reduce the cost of living WITHOUT lowering home prices.I’d add the cost of prescriptions to the list of things the admin might target in the coming months to help reduce the amount people spend every month, where the target seems “easy” from a politician’s standpoint. Picking on babies and puppies is bad for getting re-elected, but I’m not sure the same applies to insurance companies, etc. As another client told me, look for Emerging Market Populace Policies to be enacted whether you like them or not, they make sense or not, or have ever even worked! It is the nature of the beast at the moment.Bottom LineStay warm (again). I say this from California with all sincerity. I did manage to be in Palm Beach for 5 days last week and California for 9 day (this week and next) – so maybe I’m a pretty decent strategist after all.I think electricity might be a problem here for crypto, AI, and the consumer. Hence maybe why we see a bit more weakness, and it has little or nothing to do with Warsh, just the realization that some other issues are real and positioning has become very bullish (or at least it was coming into Thursday). Tyler DurdenSun, 02/01/2026 - 14:00

#CRYPTO#COMMODITIES
Enterprises are measuring the wrong part of RAG
venturebeathace 63d

Enterprises are measuring the wrong part of RAG

Enterprises have moved quickly to adopt RAG to ground LLMs in proprietary data. In practice, however, many organizations are discovering that retrieval is no longer a feature bolted onto model inference — it has become a foundational system dependency.Once AI systems are deployed to support decision-making, automate workflows or operate semi-autonomously, failures in retrieval propagate directly into business risk. Stale context, ungoverned access paths and poorly evaluated retrieval pipelines do not merely degrade answer quality; they undermine trust, compliance and operational reliability.This article reframes retrieval as infrastructure rather than application logic. It introduces a system-level model for designing retrieval platforms that support freshness, governance and evaluation as first-class architectural concerns. The goal is to help enterprise architects, AI platform leaders, and data infrastructure teams reason about retrieval systems with the same rigor historically applied to compute, networking and storage.Retrieval as infrastructure — A reference architecture illustrating how freshness, governance, and evaluation function as first-class system planes rather than embedded application logic. Conceptual diagram created by the author.Why RAG breaks down at enterprise scaleEarly RAG implementations were designed for narrow use cases: document search, internal Q&A and copilots operating within tightly scoped domains. These designs assumed relatively static corpora, predictable access patterns and human-in-the-loop oversight. Those assumptions no longer hold.Modern enterprise AI systems increasingly rely on:Continuously changing data sourcesMulti-step reasoning across domainsAgent-driven workflows that retrieve context autonomouslyRegulatory and audit requirements tied to data usageIn these environments, retrieval failures compound quickly. A single outdated index or mis-scoped access policy can cascade across multiple downstream decisions. Treating retrieval as a lightweight enhancement to inference logic obscures its growing role as a systemic risk surface.Retrieval freshness is a systems problem, not a tuning problemFreshness failures rarely originate in embedding models. They originate in the surrounding system.Most enterprise retrieval stacks struggle to answer basic operational questions:How quickly do source changes propagate into indexes?Which consumers are still querying outdated representations?What guarantees exist when data changes mid-session?In mature platforms, freshness is enforced through explicit architectural mechanisms rather than periodic rebuilds. These include event-driven reindexing, versioned embeddings and retrieval-time awareness of data staleness.Across enterprise deployments, the recurring pattern is that freshness failures rarely come from embedding quality; they emerge when source systems change continuously while indexing and embedding pipelines update asynchronously, leaving retrieval consumers unknowingly operating on stale context. Because the system still produces fluent, plausible answers, these gaps often go unnoticed until autonomous workflows depend on retrieval continuously and reliability issues surface at scale.Governance must extend into the retrieval layerMost enterprise governance models were designed for data access and model usage independently. Retrieval systems sit uncomfortably between the two.Ungoverned retrieval introduces several risks:Models accessing data outside their intended scopeSensitive fields leaking through embeddingsAgents retrieving information they are not authorized to act uponInability to reconstruct which data influenced a decisionIn retrieval-centric architectures, governance must operate at semantic boundaries rather than only at storage or API layers. This requires policy enforcement tied to queries, embeddings and downstream consumers — not just datasets.Effective retrieval governance typically includes:Domain-scoped indexes with explicit ownershipPolicy-aware retrieval APIsAudit trails linking queries to retrieved artifactsControls on cross-domain retrieval by autonomous agentsWithout these controls, retrieval systems quietly bypass safeguards that organizations assume are in place.Evaluation cannot stop at answer qualityTraditional RAG evaluation focuses on whether responses appear correct. This is insufficient for enterprise systems.Retrieval failures often manifest upstream of the final answer:Irrelevant but plausible documents retrievedMissing critical contextOverrepresentation of outdated sourcesSilent exclusion of authoritative dataAs AI systems become more autonomous, teams must evaluate retrieval as an independent subsystem. This includes measuring recall under policy constraints, monitoring freshness drift and detecting bias introduced by retrieval pathways.In production environments, evaluation tends to break once retrieval becomes autonomous rather than human-triggered. Teams continue to score answer quality on sampled prompts, but lack visibility into what was retrieved, what was missed or whether stale or unauthorized context influenced decisions. As retrieval pathways evolve dynamically in production, silent drift accumulates upstream, and by the time issues surface, failures are often misattributed to model behavior rather than the retrieval system itself.Evaluation that ignores retrieval behavior leaves organizations blind to the true causes of system failure.Control planes governing retrieval behaviorControl-plane model for enterprise retrieval systems, separating execution from governance to enable policy enforcement, auditability, and continuous evaluation. Conceptual diagram created by the author.A reference architecture: Retrieval as infrastructureA retrieval system designed for enterprise AI typically consists of five interdependent layers:Source ingestion layer: Handles structured, unstructured and streaming data with provenance tracking.Embedding and indexing layer: Supports versioning, domain isolation and controlled update propagation.Policy and governance layer: Enforces access controls, semantic boundaries, and auditability at retrieval time.Evaluation and monitoring layer: Measures freshness, recall and policy adherence independently of model output.Consumption layer: Serves humans, applications and autonomous agents with contextual constraints.This architecture treats retrieval as shared infrastructure rather than application-specific logic, enabling consistent behavior across use cases.Why retrieval determines AI reliabilityAs enterprises move toward agentic systems and long-running AI workflows, retrieval becomes the substrate on which reasoning depends. Models can only be as reliable as the context they are given.Organizations that continue to treat retrieval as a secondary concern will struggle with:Unexplained model behaviorCompliance gapsInconsistent system performanceErosion of stakeholder trustThose that elevate retrieval to an infrastructure discipline — governed, evaluated and engineered for change — gain a foundation that scales with both autonomy and risk.ConclusionRetrieval is no longer a supporting feature of enterprise AI systems. It is infrastructure.Freshness, governance and evaluation are not optional optimizations; they are prerequisites for deploying AI systems that operate reliably in real-world environments. As organizations push beyond experimental RAG deployments toward autonomous and decision-support systems, the architectural treatment of retrieval will increasingly determine success or failure.Enterprises that recognize this shift early will be better positioned to scale AI responsibly, withstand regulatory scrutiny and maintain trust as systems grow more capable — and more consequential.Varun Raj is a cloud and AI engineering executive specializing in enterprise-scale cloud modernization, AI-native architectures, and large-scale distributed systems.

#TECH
Can ChatGPT Outperform the Market? Final Report
hackernoonhace 63d

Can ChatGPT Outperform the Market? Final Report

This report evaluates a six-month live trading experiment in which a large language model (ChatGPT) managed a micro-cap equity portfolio under strict, forward-only constraints.

#TECH