# Decision Points in AI Agent Development # Temperature 🎯 The Hook Are you using the same temperature for every task in your agent? Temperature isn't just a "creativity knob." In agent systems, structured outputs, tool calls, and user-facing responses each need fundamentally different temperature settings. Using one value for everything is leaving performance on the table. 📋 Overview Temperature controls how "peaked" or "flat" the probability distribution is when an LLM selects its next token. Near 0, the highest-probability token wins almost every time, producing deterministic and stable output. Higher values flatten the distribution, allowing lower-probability tokens through, increasing diversity and creativity. In AI agent systems, the optimal temperature varies dramatically across contexts: generating structured output, assembling tool call arguments, and producing free-form text each call for different settings. Temperature should be treated as a dynamic variable that shifts with task type, not a single fixed constant. 🔍 Decision Points Temperature is primarily driven by task_variability -- how routine vs. creative the task is. The decision flow is straightforward 🧭 1. Structured output (JSON / function calling)? → 0.0-0.3 2. Accuracy-first (fact extraction, classification, summarization)? → 0.2-0.5 3. Dialogue, explanation, communication? → 0.5-0.7 4. Creative writing, brainstorming, candidate generation? → 0.7-1.0 Additionally, higher failure_cost pushes the temperature ceiling down, and high cost_sensitivity environments should account for retry cost increases from higher temperatures. 💡 Key Details Reference values by task type 📊 - Structured output (JSON / function calling): 0.0-0.3. Minimizing schema violations is the priority - Classification, extraction, data transformation: 0.0-0.2. Accuracy and reproducibility are paramount - Summarization, explanation, customer support: 0.5-0.7. Balance naturalness with accuracy - Creative writing, brainstorming, candidate generation: 0.7-1.0. Diversity is the source of value - Tool argument generation: 0.0-0.2. Precise function and argument names are non-negotiable - Planning and reasoning: 0.3-0.6. Some exploration helps, but maintain logical consistency Start structured output temperature at 0. If schema violations occur at 0, the problem is your prompt or schema -- never rely on higher temperature to "accidentally" produce correct output 🚫 ⚖️ Trade-offs Too low and conversations become robotic 🤖 The model returns identical answers to identical questions, giving users a "template response" impression. Best-of-N sampling also breaks down -- candidates become near-identical, costing N times more for essentially N=1 results. Too high and structured outputs start breaking 💥 JSON field names drift, types mismatch, hallucinations increase -- especially dangerous for proper nouns, numbers, and dates. Tool call instability and loss of reproducibility compound the problem. Monitor the retry cost impact of temperature changes. If schema violation rates exceed roughly 5%, consider lowering the temperature. 🛠️ Use Cases Vary temperature by pathway within a single agent 🔀 Planning steps at 0.3-0.5, tool argument generation at 0.0-0.2, user-facing responses at 0.5-0.7. When switching models, adjust temperature simultaneously for a natural fit. Using Best-of-N? You need to raise the temperature. Generating N=5 candidates at temperature 0 produces 5 near-identical outputs. For N>1, set temperature to 0.5-0.8 and let a Judge select the best from diverse candidates. Be careful combining temperature with top_p ⚠️ Adjusting both simultaneously creates multiplicative effects with unpredictable behavior. As a rule, tune one and leave the other at its default. #AIAgents# #SoftwareArchitecture#

CPI Details: CPI rose 0.6% MoM after rising 0.4% MoM in August; it rose 3.7% YoY The index for gasoline rose 9.1% in September, accounting for over half of the monthly all items increase. The energy index increased 5.2% over the month following a 2.1% increase in August. Electricity CPI rose 0.2% over the month. In contrast, the natural gas index decreased 0.3% in September. The shelter index rose 0.4% in September after rising 0.3% in August. The index for food was unchanged in September, as the index for food away from home increased 0.2% and the index for food at home fell 0.1%. It increased 2.5% over the last year.Three of the six major grocery store food group indexes decreased in September, while the other three indexes increased. The meats, poultry, fish, and eggs index decreased 0.5%over the month, as the beef index fell 1.0%. The index for other food at home decreased 0.2 percent in September. The dairy and related products index decreased 0.9% over the month, as the milk index declined 2.0%. The energy index increased 20.4% YoY due in large part to the index for gasoline rising 34.2% over the same period. The natural gas index increased 5.0% YoY and the electricity index rose 4.4% YoY. Core CPI rose 0.3% in September, same as August. Indexes that increased over the month include recreation, used cars and trucks, education, airline fares, medical care, and apparel. Conversely, the index for communication was among the small number of major indexes that decreased in September. Shelter CPI increased 0.4% MoM, after rising 0.3% in August.Index for owners’ equivalent rent rose 0.3% in September, and the index for rent increased 0.2%. Lodging away from home index rose 2.0% in September after rising 2.4% the previous month. Index for recreation increased 0.5% over the month after being unchanged in August. Used cars and trucks index rose 0.7% in September, and the education index increased 0.6%. New vehicles index, the personal care index, and the motor vehicle insurance index also increased over the month. Index for airline fares increased 1.2% in September, and the index for apparel rose 0.3%. The medical care index increased 0.1% in September after falling 0.2% in August. The index for hospital services rose 0.7% over the month. Prescription drugs index declined 1.2% in September. Index for physicians’ services was unchanged over the month. Communication CPI declined 0.2% in September after rising 2.3% in August. Index for household furnishings and operations was unchanged over the month.

Seednace 2.5 Prompt: Scene Understanding A raw underground streetwear performance film inside a decaying industrial building. The performers themselves do not physically transform; instead, the space, framing, distance, and visual dominance between them continuously evolve. The short begins with Performer A controlling the frame, gradually introduces Performer B through the architecture, then lets both performers exchange visual dominance until they finally unite in a strong symmetrical fashion tableau. The mood is cold, confident, mysterious and controlled rather than chaotic. Shot Design First 3 Seconds — Hook 0.0–1.2s Extreme 12mm fisheye floor-level shot directly beside Performer A’s sneaker. Performer A steps toward and almost over the camera. His black sneaker enlarges naturally from perspective until it briefly blocks most of the frame. Real foot contact, fabric inertia and subtle camera vibration. 1.2–3.0s As the shoe clears the lens, the handheld camera rapidly slides backward beneath the broken staircase. The movement reveals Performer B standing completely still in deep background beneath damaged stair rails. Performer A crosses foreground while Performer B maintains direct eye contact with camera. The spatial reveal should feel surprising but physically continuous. Middle 7–8 Seconds — Visual Pursuit 3.0–4.7s Whip-pan toward Performer B. Camera pushes aggressively into a distorted fisheye medium close-up. He performs restrained shoulder and hand gestures. Cut on beat to Performer A against cracked brick and blown-out windows. Brief stable low-angle portrait. He slowly adjusts his white sunglasses. Hold long enough to clearly preserve facial structure and reflections. 4.7–6.7s Hard cut to an ultra-wide floor-level shot across rubble. Performer A walks rhythmically toward camera. The camera retreats at almost the same speed, creating strong parallax across broken concrete and debris. He passes extremely close to one side of the lens. Natural foreground smear and fabric movement. As A clears frame, Performer B is revealed behind him already advancing through the same space. No teleportation. 6.7–8.8s Camera begins a controlled circular orbit around Performer B using a 12mm fisheye. He leans slightly toward the lens. One pale blue-grey trouser leg and white sneaker enter extreme foreground with natural perspective enlargement. The camera completes the rotation into Performer A standing several metres away. A fast push-in lands on an almost symmetrical extreme close-up of Performer A’s face and white wraparound sunglasses. Hold briefly. Real reflections of windows and ruined architecture appear in the lenses. 8.8–10.5s Pull sharply away from Performer A. Performer B appears centered between rusted staircase rails. He folds his arms, pauses, then opens them into one restrained rhythmic gesture. He begins walking toward camera while the camera retreats down a narrow debris corridor. Foreground floor textures rush past naturally. Final 4–5 Seconds — Payoff 10.5–12.2s Fast whip-pan upward into a high oblique view. Both performers are visible for the first time within the same architectural composition. Camera drops rapidly from the elevated angle toward ground level. Transition into an extreme fisheye low-angle shot. Both performers lean into the frame from opposite sides, briefly dominating the lens. No exaggerated body deformation beyond realistic fisheye distortion. 12.2–13.4s Snap to a brighter ruined window bay. Performer B enters first and stops. Performer A steps into frame beside him. They stand shoulder-to-shoulder. Minimal movement. Direct gaze toward lens. 13.4–15.0s The performers remain almost perfectly still. Camera suddenly performs a fast but controlled pull-back through the long destroyed corridor. Broken beams, rails, doorways and rubble rush past with powerful parallax. The performers rapidly become smaller but remain clearly visible together in the bright opening. Finish on a very wide architectural composition. Hard beat stop. Final Seedance 2.0 Prompt Create a 15-second photorealistic underground streetwear fashion-performance film in 16:9 inside a large abandoned industrial building constructed from exposed concrete, damaged red brick and aging timber. The building contains a broken wooden staircase, rusted stair rails, exposed beams, cracked plaster, dusty concrete floors, scattered rubble, damaged doorways, narrow debris corridors and enormous blown-out window openings flooded with natural daylight. Maintain exact visual continuity for two male performers throughout the entire film. Performer A: pale skin, backward muted-green cap, white wraparound sunglasses, small hoop earrings, oversized black-and-white abstract fleece jacket, loose dark trousers and black sneakers. Performer B: black balaclava, white oval wraparound sunglasses, teal-green work jacket, dark shirt, extremely baggy pale blue-grey jeans and clean white sneakers. Their clothing, proportions, accessories and identities must remain consistent in every shot. Use realistic skin texture, natural body proportions, believable fabric weight, realistic clothing inertia, grounded foot contact and physically plausible human movement. Visual style: raw underground European streetwear music video, documentary-real, cold, confident and mysterious. Shoot as if using a full-frame cinema camera with 12mm fisheye and 18mm ultra-wide lenses. Camera movement must feel physical and human: handheld floor-level tracking, close body passes, low-angle push-ins, pull-backs, controlled orbiting movements, short whip-pans, brief Dutch angles and occasional rotational transitions. Preserve true camera momentum, acceleration, deceleration and architectural parallax. No teleportation, instant position changes or impossible camera paths. 30 fps with a natural 1/60 shutter feel, organic motion blur, mild lens breathing, realistic fisheye edge distortion, subtle sensor noise and slight documentary exposure variation. 0–3 seconds Open with an extreme 12mm fisheye floor-level shot beside Performer A's black sneaker. He steps directly toward and almost over the lens. His shoe grows naturally in perspective until it briefly obscures most of the camera. As his foot clears the frame, the handheld camera rapidly slides backward beneath the damaged wooden staircase. Reveal Performer B standing motionless in deep background between broken stair rails while Performer A crosses close foreground. Maintain believable spatial continuity. 3–5 seconds Whip-pan toward Performer B beneath the stairs. Push rapidly into a distorted fisheye medium close-up while he performs restrained rhythmic shoulder and hand gestures. Cut sharply on a beat to Performer A in a stable low-angle portrait against cracked brick walls and bright broken windows. Performer A slowly touches and adjusts his white wraparound sunglasses. Hold briefly for stable facial detail and realistic environmental reflections in the lenses. 5–7 seconds Hard cut to a very wide floor-level view across dusty rubble. Performer A walks and moves rhythmically toward camera while the camera tracks backward. Produce strong architectural parallax and rushing debris texture. Performer A passes extremely close beside the lens, creating natural motion smear. As he clears foreground, reveal Performer B already advancing through the same corridor behind him. 7–9 seconds Perform a controlled circular 12mm fisheye orbit around Performer B. He leans slightly toward camera. One white sneaker and pale blue-grey trouser leg enter extreme foreground and enlarge naturally from perspective. Complete the rotation toward Performer A positioned farther inside the building. Execute a fast physical push-in into an extreme symmetrical close-up of Performer A. Hold his face and white sunglasses briefly with realistic skin, sharp identity preservation and natural reflections. 9–10.8 seconds Rapidly pull away. Reveal Performer B centered between damaged stair rails. He folds his arms, pauses, then opens them into a restrained hand gesture. He begins advancing toward camera. Camera retreats along the debris-covered corridor, creating strong floor-level parallax and believable spatial compression. 10.8–12.4 seconds Whip upward into a high oblique camera angle revealing both performers near the staircase. Immediately descend through a fast physical camera drop into an extreme fisheye low angle. Both performers lean naturally toward the lens from opposite sides and perform small confident gestures. Avoid excessive deformation. 12.4–15 seconds Snap to a brighter ruined bay illuminated by large blown-out windows. Performer B enters first and stops. Performer A steps beside him. They finish shoulder-to-shoulder, facing directly toward camera with minimal movement and calm confidence. On the final musical accent, perform a rapid but physically controlled pull-back down the long ruined corridor. Broken doors, beams, rails and floor debris rush past with strong realistic parallax. Both performers remain stationary together in the distant bright opening. End on an extremely wide architectural shot with both men small but clearly readable at the center of the composition. Lighting is entirely motivated by daylight entering through broken windows and door openings. Use cool cyan-green ambient shadows, rusty brown timber and brick, dirty cream plaster, subtle highlight bloom and deep but detailed blacks. Exposure should feel slightly imperfect and documentary-real, never polished studio lighting. Color grade: muted teal and green, dirty cream, rust red, mild halation, restrained contrast and subtle filmic texture. Audio: instrumental only. No vocals and no dialogue. Dark industrial hip-hop/electronic instrumental around 103 BPM using dry kick drums, tight snares, deep sub-bass pulse, sparse metallic percussion, low atmospheric drone and subtle natural room reverberation. Synchronize major push-ins, whip-pans, cuts and the final pull-back to musical accents. Add quiet realistic shoe scuffs, trouser and jacket fabric movement, distant wind and subtle abandoned-building ambience. No captions, subtitles, logos, watermarks or interface graphics. Avoid plastic skin, artificial facial smoothing, extra fingers, duplicated limbs, changing clothes, warped sunglasses, floating rubble, disappearing objects, rubbery body movement, impossible camera motion, excessive sharpening, extreme bloom or fake CGI-looking lighting. The final result should feel like a real fashion crew entered an abandoned industrial structure with an ultra-wide cinema camera and captured an aggressive but controlled underground music-video performance. Alternate Version — More Intense Make the first half much more confrontational. Performer A repeatedly passes within centimeters of the lens while Performer B appears through different layers of the staircase. Increase whip-pan velocity, lower the camera even closer to the floor and synchronize nearly every major movement to sharp percussion hits. During the final pull-back, both performers take one synchronized step toward camera just before the camera launches backward down the corridor. The performers must still remain realistic, grounded and visually consistent. Alternate Version — More Emotional Reduce the aggression. Begin with Performer A standing alone inside the huge ruined building. Introduce Performer B gradually through distant doorways and staircase openings. Use slightly longer takes, slower camera drift and more moments of stillness. The two performers eventually walk toward each other from different areas of the building and silently meet in the bright final bay. The final pull-back reveals how small they are within the enormous decaying architecture. Keep the same streetwear styling but give the film a lonely, atmospheric quality. Alternate Version — More Stylized Push the fisheye language harder while preserving photorealism. Use dramatic symmetrical compositions, exaggerated architectural vanishing points, rotating camera transitions and extremely close foreground passes. Create visual match-cuts between circular sunglasses, staircase curves and blown-out window openings. During the final shot, begin extremely close to both performers before pulling backward through multiple architectural frames until they become centered silhouettes against the daylight. Maintain realistic movement, real textures and physically plausible camera transitions.

🇺🇸US DATA RUNS HOT: INFLATION STICKY, DEMAND FIRM INFLATION • Headline PCE MoM: +0.2% vs +0.1% est.; prior -0.1% • Headline PCE YoY: +3.7% vs +3.6% est.; prior +3.7% • Core PCE MoM: +0.2% vs +0.2% est.; prior +0.1% • Core PCE YoY: +3.3% vs +3.3% est.; prior +3.3% GDP • Q2 GDP annualized: +1.5% vs +1.5% est.; prior +1.5% • Personal consumption: +3.4% vs +3.2% est.; prior +3.2% • GDP Price Index: +6.4% vs +6.2% est.; prior +6.2% • Core PCE QoQ: +3.6% vs +3.4% est.; prior +3.4% INCOME & SPENDING • Personal income MoM: +0.4% vs +0.2% est.; prior +0.2% • Personal spending MoM: +0.2% vs +0.1% est.; prior +0.3% • Real personal spending MoM: 0.0% vs 0.0% est.; prior +0.4% DURABLE GOODS • Durable goods orders MoM: +1.1% vs +0.5% est.; prior +0.5% • Durables ex-transportation: +0.4% vs +0.6% est.; prior +0.7% • Core capital goods orders: +0.2% vs +0.7% est.; prior revised to +1.7% • Core capital goods shipments: +1.4% vs +1.0% est.; prior revised to +2.4% BOTTOM LINE: Inflation remains sticky while consumer demand and headline durable goods beat expectations. GDP growth was in line, but stronger price pressures could keep the Fed cautious on rates.

📱 $META Street Takeaways - Meta Platforms Q3 Earnings ($680.20 -$17.32) Overview: • Shares (2.5%). Revenue +26.1% y/y beat by 2.0%. EPS ex-items also ahead by 13.1%/$0.86. OM ex-legal charge 340bps above consensus. CapEx $33.94B vs SA $40.95B. Free cash flow $2.12B vs SA ($9.22B). • Daily active people +3% y/y missed by (0.3%). ARPP 2.3% above. • Q4 revenue guided above at the mid. FY CapEx range narrowed toward high end of prior range. • Intends to file a restated certificate of incorporation in Delaware to establish Class C common stock, effective date of 19-Jan-27. • Analysts were positive on the results and outlook. While compute remains something of a limiting factor in Muse development, solid trends for the product were highlighted as >10M have deployed in the US and Canada with retention materially better than that seen in Threads rollout. Furthermore, shopping the primary use case viewed as positive indicator of future capacity for monetization. Significant runway seen remaining in core ads with FOA revenue +26% y/y in Q3 on 10% growth in impressions and 14% in average price per ad. Record engagement for Facebook and Instagram pointed to while MAUs for other apps are also solid. New class of shares seen enabling capital raises without diluting Mr. Zuckergberg. $META #META# #Earnings# #MetaPlatforms# #Tech# #Facebook# #instagram# #Muse# #Threads# #Ai# #SocialMedia# #EPS# #Zuckerberg# #FinTwit# #CapEx# Gain immediate access to earnings insights with #StreetAccount#:

August CPI landed exactly on consensus this morning, 3.4% headline and core ticking down to 2.4%, and that is not the rescue the headline makes it sound like. The Fed is being priced for a hike next Wednesday, not a cut. In-line data does nothing to stop one. Here is what the annual numbers are hiding. Month on month, headline went from 0.1 in July to 0.4 in August. Core went 0.2 to 0.3. A core run rate of 0.3 a month compounds to roughly 3.7% a year, which is not 2%. The twelve-month figures are easing because of what fell out of the base, and the monthly prints are doing the opposite. Warsh said at Jackson Hole two weeks ago that recent readings "do not tell me that underlying inflation trends have meaningfully improved." The underlying trend is the monthly number. He also said financial conditions do not look restrictive enough. Before that speech the market had a September hike near one in three. It has been 56 to 60 percent since. And yesterday's PPI is the piece almost nobody has paired with this one. Producer prices are running 5.4% over twelve months against 3.4% at the consumer, core 4.7% against 2.4%. In August itself the two moved identically, 0.4 headline and 0.3 core on both sides. But a month ago that headline gap was 1.4 points, not 2.0. Producer prices reaccelerated from 4.8% to 5.4% while the consumer number sat still. Both sides were pushed by the same thing. Petrol up 3.9% drove over a third of the consumer increase. Diesel up 24.1% drove over a third of the producer goods increase. Consumer energy is up 16.3% on the year, petrol 27.4%, heating oil 52.0%. The standard answer is that a central bank looks through an energy shock, because rates do not produce diesel. So look at where the diesel went in yesterday's release: truck freight up 2.0%, transportation and warehousing up 2.3%. That is the shock leaving the energy line and entering services. Once it is in freight it is in everything that moves, and it stops being the kind of thing you get to look through. The bar has quietly inverted. Data now has to prove a hike is not needed, and this print did not do that. There is nothing else of size on the calendar before Wednesday. One more thing worth sitting with. Warsh's stated complaint is that financial conditions are too loose. $SPY is up 0.62% and $QQQ 0.66% pre-market on this print. Every risk rally between now and Wednesday is another data point for his own argument. That is an uncomfortable loop to be long into. The curve is already splitting on it. $TLT up 0.27% while $IEF is down 0.13% pre-market, which is the front end pricing tighter policy and the long end pricing the slower growth that follows. $XLE down 0.55%, on a day energy drove the entire print, is the odd one out. What would change my read: strip the energy pass-through out and watch what core does. If freight and transport services cool in the September PPI while diesel stays high, the shock stayed contained and the look-through argument survives. If core services keep firming while the annual headline drifts lower, then the annual number is lagging comfort and the Fed will treat it that way. And if they do hold on Wednesday, watch whether the statement leans on energy as the reason, because that tells you they have decided this is supply rather than demand.

Premarket movers: Magnificent Seven stocks are mostly lower: Apple (AAPL) slips 1% after after Loop Capital Markets downgraded the stock to hold, citing iPhone shipment reductions (Amazon (AMZN) -0.2%, Microsoft (MSFT) +0.4%, Tesla (TSLA) -0.4%, Alphabet (GOOGL) -0.3%, Nvidia (NVDA) -0.7%, Meta Platforms (META) -0.8%, Apple (AAPL) -1%) 3M (MMM) rises 1.4% higher after topping Wall Street’s profit estimates and raising its full-year outlook for the second straight quarter. Alvotech (ALVO) jumps 7.4% after the US Food & Drug Administration approved a new manufacturing process for the drug Simlandi. Arhaus (ARHS) rises 4% after Piper Sandler upgraded the furniture company to overweight, writing that “a positive business inflection appears to be underway.” AST SpaceMobile (ASTS) falls more than 2% after Viasat petitioned the FCC to reject the company's request for additional overseas spectrum authority CoreWeave (CRWV) falls over 3% even after Cathie Wood's ARK bought the dip on Monday; the stock had rallied yesterday after Goldman upgraded it to Buy General Motors (GM) falls 7% after narrowing its full-year profit forecast. Halliburton (HAL) rises 3.7% after the oilfield services company reported third-quarter adjusted earnings per share that beat the average analyst estimate as drilling and evaluation revenue was higher than expected. The company also forecast mid-single digit revenue growth for 2027. Hertz (HTZ) climbs 7.1% after Oro Mobility, which is backed by the car rental company, secured an $80 million strategic equity investment from a consortium of investors, including Cohen Private Ventures. MSCI Inc. (MSCI) gains 6.5% after the index provider reported adjusted earnings per share for the third quarter that beat the average analyst estimate. Nebius (NBIS) drops more than 2%, tracking weakness across AI-related stocks NuScale Power (SMR) falls 7.2% after Barclays cut its recommendation on the nuclear reactor company to underweight as it sees risks from a lack of contracts. Rapport Therapeutics’ (RAPP) soars 34% after the drug developer said a mid-stage trial of its experimental therapy for acute bipolar mania met the primary endpoint. Shares have since pared gains to 28%. Solstice Advanced Materials (SOLS) gains 2.9% after the Financial Times reports that activist hedge fund Starboard Value has built a stake in the Honeywell spinoff, citing unidentified people familiar with the matter. 3M (MMM) in focus after signing a deal with Amazon that includes a warrant for 3.2 million shares, vesting up to $6 billion Analyst actions: BlackRock (BLK) raised to Buy at TD Cowen; TKO raised to Buy at Arete; Prologis (PLD) raised to Buy at Freedom Capital; Trex (TREX) raised to Buy at Stifel; Live Nation (LYV) cut to Sell at CFRA; Werner (WERN) cut to Neutral at Baird; nVent (NVT) initiated Overweight at JPMorgan

DOMINO’S PIZZA $DPZ Q3’26 EARNINGS HIGHLIGHTS 🔹 Revenue: $1.19B (Est. $1.18B) 🟢; +3.4% YoY 🔹 EPS: $4.60 (Est. $4.37) 🟢; +12.7% YoY 🔹 Total Gross Margin: 40.2% (Est. 40.3%) 🔴 Same-Store Sales: 🔹 U.S.: +0.3% 🔹 International (ex-FX): +0.2% Key Updates: 🔹 CEO Role: Joe Jordan steps into the CEO role Other Q3 Metrics: 🔹 Global Retail Sales Growth ex-FX: +3.2% 🔹 Global Net Store Growth: 175 stores 🔹 Supply Chain Gross Margin: 12.2% 🔹 Oper Income: $241.8M (Est. $230M) 🟢; +8.3% YoY 🔹 Net Income: $152.2M (Est. $145M) 🟢; +9.2% YoY 🔹 YTD Free Cash Flow: $459.8M; -7.2% YoY Capital Return: 🔹 Share Buybacks: $73.6M 🔹 Dividend: $1.99/share 🔹 Remaining Authorization: $1.15B Comments: 🔸 “Our opportunity now is to capitalize on those strengths by getting even closer to our customers, helping our franchisees grow their profitability and turning our best performance into our everyday performance.”

CHOOSE YOUR FIGHTER Warren Buffett vs Stanley Druckenmiller Buffett stepped down as Charman of Berkshire Hathaway $BRK.B today. Druckenmiller has never had a losing year in more than 30 years. Here are their full portfolios as of Q2 2026. BERKSHIRE HATHAWAY - Apple $AAPL: 22% - American Express $AXP: 17% - Google $GOOGL: about 12.6% across both share classes - Coca-Cola $KO: 11% - Bank of America $BAC: 9.2% - Chevron $CVX: 4.7% - Occidental $OXY: 4.3% - Chubb $CB: 3.9% - Moody's $MCO: 3.7% - Kraft Heinz $KHC: 2.6% - DaVita $DVA: 2.1% - Delta $DAL: 1.8% - SiriusXM $SIRI: 1.2% - VeriSign $VRSN: 0.8% - Kroger $KR: 0.7% - Liberty Live $LLYVA: about 0.6% across both share classes - Ally $ALLY: 0.4% - Lennar $LEN: 0.4% - New York Times $NYT: 0.4% - Capital One $COF: 0.2% - Louisiana-Pacific $LPX: 0.1% - Nucor $NUE: 0.1% - Macy's $M: 0.1% - NVR $NVR - Jefferies $JEF - D.R. Horton $DHI DUQUESNE FAMILY OFFICE - Natera $NTRA: 17% - Insmed $INSM: about 5.7% in shares and calls - Taiwan Semi $TSM: 5.4% - Brazil ETF $EWZ: about 5.1% in shares and calls - Amazon $AMZN: about 4.6% in shares and calls - STMicro $STM: 4.5% - S&P 500 Equal Weight $RSP: about 3.7% in calls and shares - Fox $FOXA: about 2.8% across both share classes - YPF $REPYY: 2.7% - CDW $CDW: about 2.7% in shares and calls - BBB Foods $TBBB: 2.3% - Google $GOOGL: 2.3% - Seagate blockstack:native: 2.3% - United Airlines $UAL: 2.1% - Sea $SE: 2.0% - NewAmsterdam Pharma $NAMS: 2.0% - Russell 2000 ETF $IWM: 1.9% in calls - Sandisk $SNDK: 1.5% - Revolution Medicines $RVMD: 1.4% - S&P 500 ETF: 1.3% in calls - Bitdeer: 1.2% - CRH: 1.1% - Delta: 1.1% - Tesla $TSLA: 1.0% in calls - Fluor: 1.0% - D.R. Horton: 0.9% - Coupang: 0.9% - AMD: 0.8% - Palo Alto Networks: 0.8% - Cleveland-Cliffs: 0.8% - Hut 8: 0.7% - Caris Life Sciences: 0.6% - Argentina ETF: 0.6% - Woodward: 0.5% - Meta: 0.5% in calls - Nuvation Bio: 0.5% - Protagonist Therapeutics: 0.5% - Roku: 0.5% - Cavco: 0.5% - ADMA Biologics: 0.4% - Hyperliquid Strategies: 0.4% - Rambus: 0.4% - Rhythm Pharmaceuticals: 0.4% - Champion Homes: 0.4% - Daktronics: 0.4% - PureCycle: 0.4% - Southern Copper: 0.4% - Linde: 0.4% - Entegris: 0.4% - Teva: 0.4% - Unity: 0.4% - Aeva: 0.4% - Riot Platforms: 0.4% - Qnity Electronics: 0.4% - Equinix: 0.4% - Lam Research: 0.4% - Definium Therapeutics: 0.3% - Belite Bio: 0.3% - 10x Genomics: 0.3% - Wabtec: 0.3% - Eli Lilly: 0.3% - Xenon Pharmaceuticals: 0.3% - Olema Pharmaceuticals: 0.2% - Repligen: 0.2% - Rocket Companies: 0.2% - Baidu: 0.2% - Arm: 0.2% - Carvana: 0.2% - Reddit: 0.2% - Alcoa: 0.2% - Thermo Fisher: 0.2% - Danaher: 0.2% - F5: 0.2% - Vista Energy: 0.2% - Skeena Resources: 0.2% - JBS: 0.1% - Monte Rosa Therapeutics: 0.1% - Relay Therapeutics: 0.1% - DBV Technologies: 0.1% - CCC Intelligent Solutions: 0.1% - UWM Holdings: 0.1% - Navitas Semiconductor: 0.1% - Solstice Advanced Materials: 0.1% - FTAI Aviation: 0.1% - Beam Therapeutics: 0.1% - Aurora Innovation: 0.1% - IREN: 0.1% - Grupo Financiero Galicia: 0.1% - Wave Life Sciences: under 0.1% Both own Google, Delta and D.R. Horton. Neither owns Nvidia $NVDA.