Lawrence Hecht

18K posts

Lawrence Hecht

Lawrence Hecht

@LawrenceHecht

Analyst sharing open data and sharing insight about enterprise IT, tech policy, surveys, economics. Bluesky: @lawrencehecht.info

New Jersey Katılım Nisan 2009
298 Takip Edilen3.1K Takipçiler
Lawrence Hecht retweetledi
Shubham Sharma
Shubham Sharma@mr_bumss·
@databricks Multi-model strategy is now the default → 78% of companies use two or more LLM model families → Those using 3+ jumped from 36% to 59% in four months No one's betting everything on a single vendor anymore. Model flexibility is the new AI strategy
Shubham Sharma tweet media
English
1
1
0
106
Lawrence Hecht retweetledi
Holger Zschaepitz
Holger Zschaepitz@Schuldensuehner·
Alcohol companies are sitting on huge stockpiles as demand keeps falling. Five of the world’s largest listed spirits makers – Diageo, Pernod Ricard, Campari, Brown-Forman and Rémy Cointreau – are holding about $22bn worth of ageing spirits, the highest inventory level in more than a decade. ft.com/content/9e6f02… HT @knowledge_vital
Holger Zschaepitz tweet media
English
535
1.5K
7.3K
3.4M
Lawrence Hecht retweetledi
Gergely Orosz
Gergely Orosz@GergelyOrosz·
That us engineers will not write most (or any) code by hand doesn’t mean what many replies assume it does - that there won’t be demand for SWEs. The opposite: I expect more demand for software engineers who can build reliable+complex software with LLMs! newsletter.pragmaticengineer.com/p/when-ai-writ…
Gergely Orosz tweet media
Ryan Dahl@rough__sea

This has been said a thousand times before, but allow me to add my own voice: the era of humans writing code is over. Disturbing for those of us who identify as SWEs, but no less true. That's not to say SWEs don't have work to do, but writing syntax directly is not it.

English
47
48
694
64.1K
Lawrence Hecht retweetledi
Bill Doerrfeld
Bill Doerrfeld@DoerrfeldBill·
The State of MCP report from @Zuplo has some notable nuggets... • 72% of adopters expect MCP usage to increase • 70% of devs have 2–7 MCP servers configured, often for context engineering • Security and access control remain top challenges zuplo.com/mcp-report
English
0
1
2
108
Lawrence Hecht retweetledi
Gergely Orosz
Gergely Orosz@GergelyOrosz·
From an eng responsible for AI tooling at a mid-sized company (100+ devs): "Our execs read a blog post about Claude Code and ask: 'why are we not all using it?' Me: well, none of you would approve going from $40/mo on GitHub to $65/mo on Cursor... Claude Code is $150/mo."
English
122
71
4.2K
412.4K
Lawrence Hecht
Lawrence Hecht@LawrenceHecht·
@DoerrfeldBill Perhaps they meant that among the average 10,000 person org that uses MCP servers, 15% somehow utilize an MCP server? Perhaps they are basing this on a small # of companies that are their customers, and then extrapolating the results? ...
English
1
0
0
46
Lawrence Hecht
Lawrence Hecht@LawrenceHecht·
@DoerrfeldBill I believe the local and unofficial stats, but can't trust this report unless I know how they collected the data. IT/developers don't even make up 15% of employees at large enterprises, yet the report says that's how many are running MCP servers...
English
1
0
0
57
Lawrence Hecht retweetledi
Bill Doerrfeld
Bill Doerrfeld@DoerrfeldBill·
MIT reports that 95% of genAI initiatives don't produce a successful return. So — how do you actually get ROI from an AI initiative? My feature on @CIOonline gets to the bottom of this today. cio.com/article/409515…
English
0
1
0
93
Lawrence Hecht retweetledi
Eric Jhonsa
Eric Jhonsa@EricJhonsa·
The bubble right now is in quantum computing, SMR and other gamma-squeezed retail pumps. And to some extent in select frothy large-caps. AI infra contains a few expensive names, but also a bunch going for <20x 2027 EPS if spending forecasts from major players are close to right.
English
4
1
44
5.6K
Lawrence Hecht
Lawrence Hecht@LawrenceHecht·
@CloudNativeFdn The report is hard to trust because you don't say how many people are using or even aware of the specific technologies in the charts. Can you release that information? For example, what % are aware of NVIDIA Triton and what % are actually using it?
English
1
0
0
83
CNCF
CNCF@CloudNativeFdn·
🚨 Which AI tools do 300+ cloud native developers trust for production? The new CNCF Tech Radar report with SlashData reveals the leaders across AI inference, ML orchestration, and agentic AI: NVIDIA Triton, Airflow, Metaflow, and MCP. Download now: hubs.la/Q03SBxw60 #CloudNative #AI #MLOps #CNCF
English
1
3
5
2.1K
Lawrence Hecht
Lawrence Hecht@LawrenceHecht·
@kevinroose My problem with the METR study is that they claimed to release it to the public, but then actually didn't disseminate it until approx 6 weeks after it was released. By then, attempts to criticize its methodology were too late
English
1
1
0
896
Kevin Roose
Kevin Roose@kevinroose·
This is correct, and also true of every recent AI paper (the METR slow-down study, the Apple reasoning one) that casts doubt on AI's effectiveness. People are desperate to prove that LLMs don't work, aren't useful, etc. and don't really care how good the studies are.
rohit@krishnanrohit

I think folks who know better, esp on twitter, are still underrating the extreme impact the MIT paper had about 95% of AI projects failing within enterprises. I keep hearing it over and over and over again.

English
49
31
377
171.9K
Sarbjeet Johal
Sarbjeet Johal@sarbjeetjohal·
.@CloudNativeFdn and @SlashDataHQ’s 2025 Technology Radar Report surveyed developers associated with cloud native development and the report uncovered insights on their experience and opinions on AI inference tools and engines, #AgenticAI platforms and projects, and ML orchestration tools.  Key highlights include: For AI inference tools, NVIDIA Triton, DeepSpeed, TensorFlow Serving, and BentoML were the projects that developers cumulatively placed in the adopt position. @NVIDIA Triton received the highest ratings for maturity and usefulness. Model Context Protocol (#MCP) and Llama Stack are the agentic AI projects that developer perception placed in the adopt position. Airflow and Metaflow were the two technologies that rose to the adopt position for machine learning orchestration tools. More here: prnewswire.com/news-releases/… #KubeCon @CloudNativeFdn @billyboozer @AkwyZ @Scobleizer @furrier @dvellante @RobTiffany
Sarbjeet Johal tweet media
English
1
1
4
269
Lawrence Hecht
Lawrence Hecht@LawrenceHecht·
A recent report indicates that companies want to repatriate SOME workloads to private clouds but are lack in-house skills to manage these cloud environments. NOTE, they are focused apps that have 1) security/compliance needs; 2) are data intensive; and 3) are integrated with other systems. The study did not prove that cost considerations or GenAI reqs are having a big impact on these decisions. devops.com/the-cloud-rese…
Lawrence Hecht tweet media
English
0
0
1
97
Lawrence Hecht
Lawrence Hecht@LawrenceHecht·
@GergelyOrosz @Pragmatic_Eng I appreciate all the work you did categorizing the open-ended answers. Note that Loki and Tempo are actually open source projects sponsored by Grafana, while OpenTelemetry and Prometheus are projects jused by the Grafana platform.
English
1
0
2
65
Gergely Orosz
Gergely Orosz@GergelyOrosz·
This summary feels like it snapshots the market pretty well: - Self-hosted Prometheus + Grafana is the cheapest and most work - Datadog is the most expensive but great UX, features, bells & whistles - Honeycomb amazing for high-cardinality data + traces yes
Branko@brankopetric00

Evaluated 6 observability platforms for our stack. Datadog: $8,000/month, everything included New Relic: $6,500/month, complex pricing Grafana Cloud: $2,800/month, good enough Self-hosted Prometheus: $800/month, high maintenance Elastic Stack: $4,200/month, steep learning curve Honeycomb: $5,000/month, amazing for traces We chose Grafana Cloud. Why: - Team already knew Prometheus/Grafana - 80% of features at 35% the cost - Easy migration from self-hosted - Could switch to self-hosted if needed Perfect tool doesn't exist. Good enough tool that your team will use wins.

English
27
25
587
106.1K
Lawrence Hecht retweetledi
The New Stack
The New Stack@thenewstack·
The New Stack is powered by readers like you. Take our annual Reader Survey to help guide what we cover next. It only takes 5 min, and your input means a lot. surveymonkey.com/r/2025tns?sour…
The New Stack tweet media
English
0
1
0
413
Lawrence Hecht retweetledi
The New Stack
The New Stack@thenewstack·
Got 5 min? We want to hear from you! Our annual reader survey is a quick and easy way for you to tell us what you want to read about in 2026. Help us continue to be your go-to source for cloud-native insights ➡️ surveymonkey.com/r/2025tns?sour…
The New Stack tweet media
English
0
1
0
736

Keşfet

@databricks @knowledge_vital @Zuplo @DoerrfeldBill @CIOonline @CloudNativeFdn @kevinroose @sarbjeetjohal